| Original Full Text | JOURNAL for theSCIENTIFIC STUDY of RELIGIONDiversity Dynamics: How Local Religious GroupsAppear, Persist, or Disappear over TimeJeremy SennInstitut de sciences sociales des religions (ISSR)Université de Lausanne, Lausanne, SwitzerlandJörg StolzInstitut de sciences sociales des religions (ISSR)Université de Lausanne, Lausanne, SwitzerlandChristophe MonnotFaculté de théologie protestanteUniversité de Strasbourg, Strasbourg, FranceReligious diversity is often described and measured statically. This article goes a step further by describing howcongregational religious diversity changes over time, and by exposing the mechanisms underlying these changes.We combine data from two censuses (from 2008 and 2020) of congregations in Switzerland with a sample-basednational congregation study conducted in 2008. Our main findings are the following: (1) overall, the number ofcongregations is in decline. (2) Religious diversity remains stable, but underlying features of religious diversitychange radically. For example, established Christian congregations almost only disappear while Orthodox Chris-tians see almost only new congregations. (3) Rural areas lose congregations mainly because established Christiangroups merge their parishes, while urban areas have a high turnover. (4) Some congregational characteristics suchas religious tradition, number of participants or presence of a meeting space significantly predict a congregation’slikelihood of disappearing, in contrast to other variables, including income.Keywords: congregations, religious diversity, sociology of organizations.IntroductionReligious diversity is the simultaneous presence of several religious traditions in a givenplace. When referred to at the congregational level, it is often described and measured in a staticway, with researchers seeing the existence and attributes of different religious groups as interest-ing per se. Such studies have greatly increased our knowledge by mapping diversity and by de-scribing, often at a deep level, the attributes, practices, and interrelations of local religious groupsand their context. These studies have been conducted, for example, in Germany (Becci, Bur-chardt, and Giorda 2016; Körs 2018; Krech 2009; Rebenstorf 2018), Italy (Becci, Burchardt, andGiorda 2016; Giordan 2018; Pace 2018), Spain (Fons Duocastella, Luque Capellas, and FortezaGonzales 2012; Martinez-Arino 2018), Denmark (Qvortrup Fibiger 2009; Vejrup Nielsen 2018),Switzerland (Baumann 2012;Monnot and Stolz 2014), Finland (Martikainen 2004), Great Britain(Knott 2015), Australia (Bouma 1997), and the United States (Eck 2001). A pervasive claim thatAcknowledgments: Switzerland’s changing religious diversity. The National Congregations Study Switzerland II –100015_192499Open access funding provided by Universite de Lausanne.Correspondence should be addressed to Jeremy Senn, Institut de sciences sociales des religions (ISSR), Bâtiment Anthro-pole, bureau 5066 1015, Lausanne, Switzerland. E-mail: jeremy.senn@unil.chJournal for the Scientific Study of Religion (2024) 00(0):1–24© 2024 The Authors. Journal for the Scientific Study of Religion published byWiley Periodicals LLC on behalf of Societyfor the Scientific Study of Religion.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribu-tion and reproduction in any medium, provided the original work is properly cited.2 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONthese studies make is that religious diversity is growing strongly in western societies at a congre-gational level. For example, Eck (2001) writes that the growing number of so-called “immigrantcongregations” has made the United States the “most religiously diverse nation” on earth andChaves (2018:213) summarizes a book on congregational diversity in Europe by saying that “themost prominent point of similarity” across the countries or regions covered is the growth in reli-gious diversity. Remarkably, though, many of these studies only assume that religious diversityhas grown; they do not really demonstrate it. Since they use one-wave or retrospective designs,they cannot investigate changes in diversity, and nor can they give us answers to other interestingquestions: How many congregations have appeared, persisted, or disappeared in a specific times-pan? How do different religious traditions differ with respect to the likelihood that congregationsbelonging to them will appear or disappear? How do these dynamics differ with respect to therural/urban distinction, and what determinants lead to these outcomes? How, finally, do these un-derlying dynamics translate into changes in aggregate diversity measures such as the Herfindahl–Hirschman Index (HHI)? It is such questions that we address in this article.1 We contribute to thestate of the art by conducting a longitudinal investigation of changes in religious diversity at thecongregational level, across all religions, and for a whole country, namely Switzerland. Our keyquestions are: (1) At the aggregate level, how have the number of religious congregations andcongregational diversity changed in Switzerland between 2008/2009 and 2020/20201?2 (2) Atthe underlying level, how many congregations have emerged, persisted, and disappeared withinthe different religious traditions? (3) What are the congregational characteristics that account forthe likelihood of a congregation appearing, persisting or disappearing?TheoryStability and Change with regard to Aggregate Congregational DiversityCongregational religious diversity can be defined as the existence of local religious groupsbelonging to different religious traditions in a given area (Monnot and Stolz 2014). Using indi-cators such as the HHI, it is possible to measure the evolution of diversity on an aggregate levelin a region. This aggregate change is the result of how congregations within different traditionsappear, persist and disappear over time on an underlying level—processes we call “diversity dy-namics”. However, describing the aggregate trend of religious diversity does not allow us to inferwhat has happened at the underlying level. For example, if all the congregations in one tradi-tion disappear and are replaced by an equal number of congregations belonging to a previouslyabsent tradition, the HHI will be maintained at the exact same level. The same would be true ifone group had shrunk and another grown proportionally. Similarly, a decrease in the number ofcongregations of one denomination may in some cases increase and in some cases decrease over-all diversity. We therefore study these two levels of analysis separately, starting by assessing thechange in aggregate diversity, and then examining the different forms of underlying diversity dy-namics. Differentiating these underlying and aggregate levels of religious diversity also helps usidentify a logical flaw that sometimes appears in research on religious diversity. Since most suchresearch has been cross-sectional, it has been common to investigate when congregations werefounded. This has then sometimes led to the impression that the number of new (and especiallynon-Christian) congregations has risen over time, and to the conclusion that religious diversity1Note that this question of how (dis-)appearance and persistence translate into aggregate diversity is rather different fromthe debate launched by scholars close to religious economics about the causal effect of religious diversity on aggregatereligiosity. See for this latter discussion, for example: (Finke and Stark 2005; Iannaccone 1995; Olson et al. 2020).2The first census took place between 2008 and 2009, while the second took place between 2020 and 2021. For reasons ofsimplicity, we only refer to the years “2008” and “2020”. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 3has therefore greatly increased. But such a retrospective conclusion is hasty, since it omits con-gregations that have disappeared (Fons Duocastella, Luque Capellas, and Forteza Gonzales 2012;Martinez-Arino 2018).Forms of Underlying Diversity Dynamics and Their DeterminantsFor our purposes, the appearance of a new congregation is defined as the creation of a localreligious group covered by our definition (see “Data” subsection). Thus, a religious group that isrenamed is not a new congregation, but a merger that uses a new name for the merged groups is.A new congregation can appear in different ways: it may be created spontaneously “bottom-up,”or it may be planted by the mother church “top-down”; it may be linked to immigration or not; itmay be the creation mainly of one person or of a core leadership group; it may follow a church-planting formula, or it may be the result of improvised strategies. The literature describes threefactors as being important for the appearance of new congregations. First, religious tradition: re-ligious traditions differ in how likely they are to create new congregations and how they go aboutdoing so. Some religious traditions have the explicit goal of creating as many new congregationsas possible (e.g., the emphasis on “church planting” in Evangelicalism (Hutchinson and Wolffe2012; Murphree 2018)), while others generally have neither an evangelizing nor a congregation-planting agenda (e.g., in the Jewish tradition (Harrelson 2005)). Second, immigration: whilea whole congregation may emigrate to another country, it is normally the other way around:individuals migrate for economic, political, or family reasons, and, once they find themselvesin the host country, they come together to create local groups to cater for their religious needs(Baumann 2002). This is how local Hindu, Sikh, and Muslim groups have come into existencein all western countries since the 1970s. Third, geography, with new congregations of immigrantreligions being especially common in large cities (Krech 2009), the simple reason being thatimmigrants tend to live in large cities. The disappearance of a congregation can be defined asthe moment when the local religious group ceases its regular activities. Anderson et al. (2008)have estimated that 1 percent of congregations close annually in the United States; Dougherty,Maier, and Lugt (2008) have obtained annual closure rates of 0.7 percent and 1.2 percent fortwo Christian denominations over a similar time period; and May (2018) has found a rate of 1percent for a Baptist denomination during the 1984–2011 period. These figures are surprisinglylow when compared to the rates at which other types of organization disappear (Anderson et al.2008; Dougherty, Maier, and Lugt 2008). Anderson et al. offer a compelling explanation for thislow mortality by describing congregations as “minimalist organizations” that can, in difficulttimes, rely on the commitment of their members, who may provide free services or increasetheir financial giving. Congregations can also draw on savings or endowments accumulated inmore prosperous times, sell or lease assets, or even solicit financial support from denomina-tions (2008:326). The literature also identified determinants with regard to the disappearanceof congregations, notably congregational age and size. For example, Dougherty and his col-leagues show that congregation size and age are negatively correlated with congregation closure(2008).Finally, a congregational merger may be defined as the organizational union of two or morecongregations. Mergers may take one of two forms. Either one congregation absorbs one or sev-eral other congregations, the latter thus losing their identity, or two or more congregations mergeto form a completely new congregation (with all the former congregations losing their formeridentity). The goal behindmerging organizations is normally to increase organizational efficiency.Concerning congregations, it is argued that a merger will lead to fewer people being needed fororganizational boards, to reduced costs for renting facilities, and to more participants in church(Friederich 2017; Steiner and Notter 2021; Tomberlin and Bird 2012). But there are also possibledisadvantages involved in merging congregations: for example, it may prove difficult to merge thedifferent congregational cultures, the expected economies of scale could not materialize, and the 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License4 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONquality of services provided mays deteriorate because the congregation tries to cater to a largerterritory. According to the “religious economies” model by Stark and Finke, it is especially “low-tension bodies” (denominations or congregations) that disappear via mergers (2000:206). In theframework of this theory, low-tension bodies emerge through a sect-to-church transition fromhigh-tension bodies. Once they are low-tension, they lose members and create an over-supplythat has to be corrected by mergers.MethodologyDataWe combine data from two censuses of congregations across all religious traditions, and asample-based national congregation study (NCS) conducted in 2008. Census data enable us tomeasure changes in aggregate diversity, as well as to examine underlying diversity dynamics.The effect of congregational characteristics on the latter are assessed using both census and NCSdata.2008 and 2020 CensusesThe censuses took place in 2008/2009 and 2020/2021 and used a similar methodology.The starting point for each was the well-known definition of a congregation coined by Chaves(2004:1):a social institution in which individuals who are not all religious specialists gather in physical proximity to oneanother, frequently and at regularly scheduled intervals, for activities and events with explicitly religious contentand purpose, and in which there is continuity over time in the individuals who gather, the location of the gathering,and the nature of the activities and events at each gathering.Applying this definition, we systematically counted and categorized all local religious groups inthe country (for a discussion of how to apply these criteria, see Monnot and Stolz 2018b). Wecombined all available sources of information, including existing lists of local religious groupsproduced by churches and religious federations; existing lists (published or not) compiled byscholars; existing lists on institutional websites, directories, or databases; and interviews with in-formed individuals within the religious milieus. All this information was combined and reviewedto identify local religious congregations. A congregation was retained on the final list only ifit appeared on two independent sources of information. We started out in 2020 with the 2008list and updated it systematically using all the steps already described, thus noting whether thecongregations that existed in 2008 still existed or had disappeared by 2020, and whether newcongregations belonging to that religious tradition appeared sometime between 2008 and 2020.NCS 2008The Swiss National Study of Congregations I enabled us to assess the effect of congregationcharacteristics on their likelihood of appearing, disappearing and merging. It was modeled on theNCS conducted in the United States (Monnot and Stolz 2014; Stolz and Chaves 2017). For everyrandomly selected congregation, we interviewed by telephone one key informant (in most cases,the spiritual leader3) in 2008/2009 in one of the three main languages of Switzerland, namelyGerman, French, and Italian). Approximately 250 questions focused on concrete and verifiable3Pre-tests showed that spiritual leaders were the most knowledgeable about congregational life and practices. We havealways tried to arrange a meeting with a spiritual rather than an administrative leader (e.g., the priest rather than the parishpresident in the case of a Catholic congregation). Some groups, for example, in some neo-Hindu congregations, have no 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 5congregational practices, as well as on the tangible characteristics of the organization that therespondent could provide reliable information on. The response rate was 71.8 percent. Smallreligious traditions were oversampled.Measuring Change in Aggregate Religious DiversityWe used the census data to measure the change in aggregate religious diversity, which wehave operationalized using 1 – HHI (Krech 2009; Voas, Olson, and Crockett 2002). Religiousdiversity is then computed as1 −∑Ni=1s2iwhere s is the proportion of congregations that belong to the religious tradition i, andN is the num-ber of religious traditions. We computed this formula on two levels: first for Switzerland as awhole, and then separately for each of the three categories of our size of the community vari-able, which we detail below. Obviously, the score obtained depends significantly on the religiousclassification used, as well as on the area observed. To ensure robustness, we therefore computethe diversity scores with different religious classifications (with 6, 12, 16, and 35 levels, see thecoding in Appendix A3) and control for urban/rural and language region.Measuring Underlying Diversity Dynamics and Their DeterminantsFor each listed congregation, the census data show whether it disappeared, merged, persistedor appeared between the two waves. This resulted in three dependent variables: appeared, disap-peared, and merged. These three variables indicate whether the designated event occurred for thecongregation in question between 2008 and 2020 (1= yes, 0= no).4 The census data also includeother variables that have enabled us to explore the relationship between these diversity dynam-ics and three major congregational features, from which independent variables were obtained: thereligious tradition (1), the community size (2), and the migratory nature of the congregation (3).(1) Religious tradition is mainly operationalized with a 12-level religious tradition vari-able, which distinguishes (Roman) Catholic, Reformed Protestant, classical Evangelical,charismatic Evangelical, conservative Evangelical, Orthodox Christian, other Christian,Jewish, Muslim, Buddhist, Hindu/Sikh, and other religions.5 The categories do not havethe same level of precision, since the different religions, traditions and denominationsvary greatly in the number of congregations that they have. We also used a 3-level re-ligious tradition variable to distinguish established Christian (Reformed Protestant, Ro-man Catholic, and Christ-Catholic), nonestablished Christian (all other Christians), andnon-Christian.(2) Second, the size of the community where the congregation is located wasmeasured usinga 4-level variable, based on Swiss government data. The levels are 0–1999, 2000–9999,formal spiritual or administrative leader. In such cases, we have asked for an interview with someone who knows thecongregation very well.4This split into three dichotomous coding variables is used in regression models. In the graphs, we use on the one handthe variable status, which indicates whether the congregation has disappeared since 2008, remained between 2008 and2020, or been founded since 2008, and on the other hand the dichotomous variable merged.5In this coding, Catholics are exclusively Roman Catholics. Christ Catholics, a very small but established group, arecoded as “Other Christians”. We have grouped the Sikhs with the Hindus for practical reasons, as the former have onlyfour congregations in our census data set. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License6 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION10,000–99,999, and 100,000+ inhabitants. This variable was used as a proxy for theextent to which the location of the congregation is urban or rural.(3) Third, the immigrant variable indicates whether a congregation is considered immigrant(1= yes, 0= no). We have classified congregations as immigrant if they meet one or twoof the following criteria: (1) it primarily uses a language that is not a national language6and/or (2) it organizes some of its activities specifically for people from a specific migra-tory group (e.g., Turkish–German language classes). By completing the census, we wereable to code the variable for a vast majority of both waves’ congregations (96 percent),most notably thanks to the congregations’ and denominations’ websites. This approachis not perfect, as websites do not always give absolute certainty regarding our criteria.The proportion of missing values of the variable is also unevenly distributed accord-ing to the tradition in question, which could lead to bias. We are confident, however,that our procedure provides a reasonable approximation of reality thanks to the relativecompleteness of the information available on the websites. Detailed information on thedistribution of the immigrant variable according to the tradition can be found in Ap-pendix A4. By linking census data with the NCS data set, we were also able to test theeffect of several other variables on the likelihood of the 2008 congregations to disappearand merge. Namely, we used the variable year of foundation that measures in which yearthe congregation was founded. We additionally used a dichotomous meeting space vari-able, which indicates whether the congregation has one or several places open during theweek where members can meet regularly, such as a cafeteria, a library, a games room,and so on (1 = yes, 0 = no). We also used a conflict variable indicating whether thecongregation’s respondent had reported a “conflict which required a special meeting tobe resolved” in the last 2 years (1 = yes, 0 = no). Finally, the following variables al-lowed us to assess the relationships between different types of congregational resourceson the one hand, and the disappeared and merged variables on the other: (1) ownershipof building is a 3-level variable measuring whether the congregation is the owner of thebuilding where it organizes worship. The question reads: “Does your community or fed-eration own your place of worship, can you use the place for free, or do you pay rent?”The variable is transformed into two dummy variables, with “using the place for free”being the reference category. (2) Income was measured by asking the key informant tospecify the total amount of congregational income from all sources during the past year.This variable was logged (log 10) in our analyses. (3) Attendance and membership wereoperationalized by asking the key informant how many persons participated regularly inthe religious life of the congregation, and how many people were formally members ofthe congregation. The resulting Number regular participants and Number total memberswere logged (log 10) in our analyses.Table 1 gives descriptive information on our dependent and independent variables. For ourmodeling, we use straightforward logistic regression. We used R (version 4.1.1). For replicationpurposes, the data and scripts are available upon request. They will be made available in thefollowing data repositories: FORS (https://forscenter.ch) and ARDA (https://thearda.com).6The Swiss national languages are German, French, Italian, and Romansh. If a congregation primarily uses one of thenational languages, but in a region where this language is not one of the official languages (e.g., Italian in Geneva), it isclassified as immigrant. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 7Table 1: Descriptive information on the census and the NCS 1 data setsCensus Data Set NCS 1 Data Set2008 2020 20081 - Herfindahl Index (Trad12) 0.83 0.84 –appeared (mean, SD) 0.130 (0.337) –disappeared (mean, SD) 0.193 (0.395) 0.154 (0.361)merged (mean, SD) 0.052 (0.221) 0.0391 (0.194)Immigrant 1126 (17.8%) 1227 (20.9%) –Catholic 1960 (30.9%) 1818 (30.9%) 252 (24.7%)Reformed Protestant 1073 (16.9%) 928 (15.8%) 181 (17.7%)Evangelical (classical) 561 (8.8%) 540 (9.2%) 140 (13.7%)Evangelical (charismatic) 720 (11.4%) 725 (12.3%) 88 (8.6%)Evangelical (conservative) 195 (3.1%) 128 (2.2%) 21 (2.1%)Orthodox Christian 70 (1.1%) 118 (2%) 21 (2.1%)Other Christians 775 (12.2%) 683 (11.6%) 99 (9.7%)Jewish 34 (0.5%) 29 (0.5%) 18 (1.8%)Muslim 361 (5.7%) 363 (6.2%) 64 (6.3%)Buddhist 141 (2.2%) 161 (2.7%) 47 (4.6%)Hindus/Sikhs 185 (2.9%) 174 (3%) 32 (3.1%)Others 266 (4.2%) 216 (3.7%) 59 (5.8%)Community size0–1999 1079 (17%) 943 (16%) 218 (21.3%)2000–9999 2080 (32.8%) 1909 (32.4%) 353 (34.5%)10,000–99,999 2119 (33.4%) 2065 (35.1%) 293 (28.7%)100,000+ 1063 (16.8%) 966 (16.4%) 158 (15.5%)Year of foundation (median, IQR) – – 1919 (287)Number regular participants – – 70 (120)Number total members – – 151 (481)Renting — – 250 (24.5%)Owner – – 640 (62.6%)Meeting space – – 499 (48.8%)Income (median, IQR) – – 173,437(550,000)N 6341 5883 1022ResultsStable Aggregate Diversity and Declining Number of CongregationsThere was no increase in religious diversity on an aggregate level in Switzerland between2008 and 2020. Rather, as shown in Figure 1(a), religious diversity remained stable. This is truewhen analyzing religious diversity in Switzerland generally and when controlling for urban/ruralor language region. It is also robust when varying the “resolution” of our measurement of religiousdiversity, that is, whether we use a religious tradition variable with 6, 12, 16, or 35 categories(see Appendix A3). The finding of stable religious diversity over a period of more than 10 yearsis noteworthy, since it contrasts with many studies that report or suggest increasing religiousdiversity, especially in urban contexts (Monnot and Stolz 2018a). 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License8 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONFigure 1Religious diversity (computed with the 12-level religious tradition variable) and number ofreligious congregations by urban/rural distinctions in 2008 and 2020Figure 2Number of appeared, stayed, and disappeared congregations, by religious traditionAlthough religious diversity remained remarkably stable, the number of religious congre-gations in Switzerland declined in the period observed, from 6341 to 5883 (Figure 1 (b)). Thisdecline (−7.2 percent) is the result of the fact that more congregations disappeared (1224) thanappeared (766), with 5117 persisting. Since the population grew during the same period from7,508,700 to 8,680,890, the number of inhabitants per congregation rose quite steeply from 1184in 2008 to 1476 in 2020. The secularization of Switzerland, which can be observed at the individ-ual level (Monnot and Wernli 2023; Stolz and Senn 2022), is thus accompanied by a decline inthe number of local religious communities, a decline that can be found for established Christians(−297), nonestablished Christians (−117), and non-Christians (−44). 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 9Diversity Dynamics: New Congregations, Congregations that Persist, and DefunctCongregationsAlthough religious diversity is very stable at an aggregate level, the percentage of new,persisting, and defunct congregations varies strongly between religious traditions at an under-lying level (Figure 2). Catholics, Reformed Protestants, and conservative Evangelicals reportalmost only congregations that have disappeared; Orthodox Christians report almost only newcongregations; and all others report varying percentages of turnover with both new and defunctcongregations. The highest turnover can be found among charismatic Evangelicals. As fornet growth (appearances – disappearances), all denominations have declined except OrthodoxChristians (+48 congregations), Buddhists (+20 congregations), charismatic Evangelicals (+5congregations), and Muslims (+2 congregations). Several points are worth noting with regard tothese findings. First, established groups (Catholics and Reformed Protestants) are basically in adefensive position: they are subject to many congregational disappearances, but have no newlyforming groups to counteract this decline. There is an extremely low number of new congrega-tions (seven among Catholics and 17 among Reformed Protestants), and all of these are withoutexception due to mergers where two or more congregations disappeared to create a new congre-gation with a new name. Second, the different Evangelical streams have very different dynamicsfrom one another. Conservative Evangelicals are clearly in decline, while traditional and charis-matic Evangelicals are relatively stable, although the latter have a much higher turnover. Third,Orthodox Christians have the highest growth rate with +69 percent (+48 congregations, whichmeans an annual growth of +5.7 percent). This rate is even higher than that reported for the sametradition by Giordan (2018) in his Italian mapping study. Fourth, the number of Muslim groupshas only grown very slightly (+2 groups, +0.6 percent overall). This is interesting, since theMuslim population increased by about 28 percent between 2010 and 2020 (OFS 2022). This lackof correspondence is probably related to the difficulties that Muslims face when it comes to open-ing a mosque or a prayer hall. Indeed, NCS data tell us that 22.2 percent of Muslim congregationshave experienced an “opposition problem” regarding “building or assigning a place of worship,”which is more than twice the overall mean (11 percent). An ethnography by Monnot (2016) alsoshows that Swiss Muslim associations face a considerable degree of local opposition when itcomes to renting premises for prayer or transforming a space into a prayer room. Finally, despitethe changes we describe here in terms of disappearances and appearances, the censuses also reveala considerable degree of stability: only 19.3 percent7 of congregations present in 2008 have sincedisappeared, while only 8.7 percent of congregations observed in 2020 have appeared since 2008.Figure 2 also helps us to see the limitations of congregational diversity studies based on asingle-wave design. They only allow for an assessment of past religiosity in retrospect (generallyusing congregations’ founding year). The problem with such an approach is that it involves akind of survivor bias: one cannot ask the disappeared congregations when they were founded. Ina pinch, this would not be a problem for measuring past diversity if congregations affiliated withminority traditions had the same probability of disappearing each year as majority congregations.However, as can be glimpsed from the graph,minority traditions often have a highermortality rate.By way of illustration, 17 percent of Christian congregations present in 2008 have disappeared,with this percentage being almost twice as high among non-Christian congregations (32 percent).Under these conditions, the religious diversity of the past can be underestimated by such studies.7This proportion translates to an annual mortality rate of 1.6 percent, much higher than the 1 percent annual mortalityrate reported for American congregations over the 1998–2005 period (Anderson et al. 2008). This difference cannotbe attributed to the fact that Switzerland has “established” denominations. The Swiss rate is markedly higher for non-Christian congregations (2.6 percent) and nonestablished Christians (2.2 percent), whereas it is almost identical to theAmerican rate for established (in Switzerland: mainline) Christian congregations (0.9 percent). 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License10 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONFigure 3Number of appeared and disappeared congregation, by community sizeDiversity Dynamics in Rural and Urban AreasDiversity dynamics differ in urban and rural areas. As can be seen in Figure 3, the main pro-cess in rural areas (0–1999 inhabitants) is disappearance. The disappeared rural congregations aremainly those of established Christians. This is partly because there are far fewer nonestablishedChristians and non-Christians in rural areas in the first place. The main process in urban areas,on the other hand, is turnover (especially among nonestablished Christians and non-Christians),with many congregations both appearing and disappearing. Overall, however, the number of con-gregations is also declining in urban areas, in all three categories of the 3-level religious traditionvariable (established Christians, nonestablished Christians, non-Christians).One important point regarding the disappearance of congregations in rural areas is that thisdoes not normally mean that entire villages are then without a congregation, but rather that theprovision of religious goods in rural areas is “thinned out”. As we will show below, it is a fact thatestablished congregations disappear only through merger, meaning that a merged congregationtakes over the administration of salvation goods for a village that has lost its “own” parish orcongregation. This is not possible without scaling down the provision of goods. Thus, the churchbuilding will often continue to exist, but it will only host religious services occasionally. A criticalpoint to note regarding turnover in the large cities is that it helps us change the perspective fromwhat we have learnt frommapping studies.Mapping studies havemostly shown the large diversityand number of new religious groups in large cities, sometimes interpreting this as pointing togrowth of religious diversity and even of religious activity in general. In contrast, our methodsuggests that the high number of new congregations, especially among nonestablished Christiansand non-Christians, is counterbalanced and even surpassed by the number of congregations thatdisappear. In other words, what is important in the urban context is not growth but turnover.Diversity Dynamics and ImmigrationOne major question with regard to new congregations is whether they have come about as aresult of autochthonous congregational vitality or immigration. Our data allow us to investigate 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 11Figure 4Number of congregations founded since 2008 that are immigrant and domestic, by religioustraditionthis question empirically. As seen in Table 1, the proportion of “immigrant” congregations in-creased from 2008 (17.8 percent) to 2020 (20.9 percent). When it comes to the question of howmany of the new congregations in each religious tradition are the result of immigration, the resultsdisplayed in Figure 4 partly confirm and partly confound expectations.As expected, immigration is themajor reason for the strong growth in the number ofOrthodoxcongregations in Switzerland, with 98.1 percent of new Orthodox congregations being “immi-grant congregations,” the majority of which can be assigned to national denominations (Russian,Bulgarian, Ethiopian, Romanian, etc.). Immigration is also of great importance for Muslims andcharismatic Evangelicals, for whom 75.3 percent and 49.1 percent of new congregations are“immigrant congregations”. No less surprisingly, there are no “immigrant congregations” amongthe rare newcomers within the established Christian congregations. A more unexpected result isthe low proportion of “immigrant congregations” among the new Buddhist (6.8 percent) and theHindu/Sikhs (20 percent) congregations. The appearance of new Buddhist and Hindu/Sikhs con-gregations was only marginally caused by immigration during the observation period. Those re-ligious traditions have developed an autochthonous existence that is able to generate new groups.Diversity Dynamics and MergersHowmany of the disappearances of local religious groups are due to mergers, and which con-gregations merge? Our data enable us to answer these questions for the first time for a Europeancountry across all religious traditions, and our findings are striking (Figure 5). Established de-nominations (Reformed Protestant, Catholic) disappear exclusively as a result of mergers, whilenonestablished groups hardly ever merge. In the observation period, 149 Catholic and 162 Re-formed Protestant congregations disappeared, with all merging. In contrast, all the other traditionstogether saw 913 groups disappear, with only 17 (1.9 percent) merging. Our explanation is thatthere is public recognition for Catholic and Reformed churches, which means that they exist (in 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License12 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONFigure 5Reasons for the disappearance of congregations, by religious traditionmost cantons) as institutions of public law and are thus seen as providers of public welfare8. Italso means (again, in most cantons) that their administrative structure is relatively significant,and that they receive public funds, either through church tax or through direct state funding. Asproviders of public welfare, they can evade simple closure if Catholic or Reformed members stillexist in a given territory. On the other hand, it seems much easier to justify merging two or morecongregations and then arguing that the newly merged congregation caters for the needs of thoseCatholic and Reformed members remaining. Such a merger also means that the physical churchas a spiritual building can be kept in each village. In fact, even though people may never attendchurch service, they often want their church building to remain in the village at all costs (Sahli andWüthrich 2007). Finally, a merger reduces the number of paid and voluntary staff necessary tomaintain a working congregation. Merging with another congregation is not easy and is often metwith resistance from parish members. It is therefore often cantonal authorities or their denomi-nations that encourage congregations to merge. For example, the main impulse behind the mega-merger of parishes in the city of Zurich (which merged 32 reformed parishes together) came fromthe cantonal authorities. In addition to being able to issue injunctions to merge parishes, cantonalauthorities often assist parishes in the merger process, with regard both to its planning and imple-mentation, and to its financing (Steiner and Notter 2021). According to the survey by Steiner andNotter (2021), it is also often the cantonal mother church that initiates and helps to implementmergers. Another point worth noting is that there are more Reformed Protestant than Catholicmergers in absolute terms. In relative terms, this difference is even more pronounced: twice asmany Reformed Protestant groups (15.1 percent) merged than Catholic groups (7.6 percent). Thereason for this difference is that, to avoid “hard” mergers, Catholics resort to so-called “pastoralclustering”. This is when two or more parishes pool their resources, and particularly their pas-toral staff (Berchier 2012; Monnot 2013). Thus, while each of the merged congregations retains8For a detailed analysis of the relationship between welfare and religion in Europe, see (Bäckström et al. 2016). 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 13its legal and administrative independence, the congregations together share a single pastoral teamand usually rotate the place of worship.9Thus, what is happening in Reformed and Catholic parishes actually differs little: both poolresources, and both have congregations that increasingly work together to cope with a declinein membership and resources while nonetheless trying to provide a comprehensive service in agiven territory. But, while the Reformed tradition tends tomerge parishes completely, the Catholictradition keeps the parishes as separate units, while merging their resources (such as pastoraland administrative staff). Finally, it should be noted that the greater proportion of mergers inestablished churches is in line with the theoretical proposition on this subject in the “religiouseconomies” model by Stark and Finke (2000:206).Determinants of Diversity DynamicsIn what follows, we use straightforward logistic regressions to investigate how strong therelationships are between independent variables and our three dichotomous dependent variables(appeared, disappeared, and merged). In practical terms, we use a generalized linear model witha log-link, as well as nested models, gradually introducing different explanatory variables, andfinally the 12-level religious tradition variable.We only highlight here themost important findingsthat these analyses yield. Overall, we find that using a multivariate perspective confirms what wehave already seen by way of the graphs.Table 2 presents nine logistic regression models with the dependent variables appeared, (1= appeared, 0 = otherwise), disappeared, (1 = disappeared, 0 = otherwise), and merged (1 =merged, 0 = otherwise). The population observed in the three models for the appeared variableconsists only of congregations that either appeared or persisted between 2008 and 2020 (i.e., the2020 population). The population observed in the models for the disappeared and merged vari-ables comprises only congregations that either persisted or disappeared between 2008 and 2020(i.e., the 2008 population). In the first model for the appeared dependent variable, we only in-troduce the dichotomous variable measuring whether the congregation observed was coded as“immigrant”. Among the congregations that existed in 2020, the odds of an immigrant congrega-tion having been founded since 2008 are four times higher than those for a nonimmigrant group.In the next model (Appeared 2), we introduce the 4-level variable size of community (with 0–1999 inhabitants as the reference category). We find that urban congregations that exist in 2020have twice the odds to have been founded since 2008 than rural congregations. In the third model(Appeared 3), we introduce the 12-level religious tradition variable (with Evangelical Classicalas the reference category). All non-Christian traditions, except Jewish, have a higher probabilityof founding new communities than classical Evangelicals, with the odds ratios particularly lowfor established Christians, and high for Orthodox Christians or charismatic Evangelicals.Despite taking religious tradition into account, the odds ratio for immigrant congregations istwice as high as for nonimmigrant congregations. One interesting detail is that the effect of theurban context changes from larger than 1 to smaller than 1 when introducing religious tradition,which now suggests that the effect of the urban context is to reduce the likelihood that a newcongregation will appear (odds ratio = 0.391). The reason for this is that most non-Christiangroups can be found in urban contexts; controlling for religious tradition can thus change thedirection of the rural/urban effect.9As a further explanation, the greater reluctance of Catholics to merge could be interpreted as a difference in the impor-tance, or even the sacredness, that they attach to their church buildings. However, our data do not allow take a definitivestatement. On the other hand, a study by Village and Francis (2021) shows that among English churchgoers, it is theCatholics who attach the most importance to their local religious buildings, both in terms of faith and identity. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License14 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONTable2:Oddsratioswithregardtoacongregationappearing,disappearing,orbeingsubjecttomerger(censusdata)Appeared1Appeared2Appeared3Disappeared1Disappeared2Disappeared3Merged1Merged2Merged3PredictorsOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosIntercept0.098***0.059***0.222***0.213***0.200***0.256***0.066***0.119***0.013***Migrant3.562***2.886***1.977***1.1220.9990.9410.218***0.279***0.567*2000–9999inh.1.445*0.615*0.9680.727**0.539***0.695*10,000–99,999inh.2.199***0.513***1.0640.550***0.256***0.566**100,000+inh.2.143***0.391***1.523***0.672**0.622**1.932**Catholics0.023***0.427***7.862***ReformedProtestant0.121***0.90015.702***EvangelicalCharismatic3.165***2.450***0.436EvangelicalConservative0.058**2.826***0.000OrthodoxChristian3.958***0.6240.000OtherChristians1.436*2.124***0.241Jewish0.2860.9481.948Muslim1.3931.743**0.000Buddhist6.928***2.363***0.000Hindu/Sikh4.573***4.277***0.454Others2.408***3.484***0.290Observations573357335733610261026102610261026102R2Tjur0.0440.0490.1920.0000.0040.0680.0070.0190.078Deviance4118.8444082.4773225.0085725.6165703.7185314.5202478.3102413.1092087.790AIC4122.8444092.4773257.0085729.6165713.7185346.5202482.3102423.1092119.790∗ p<.05;∗∗p<.01;∗∗∗ p<.001. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 15Table3:Oddsratioswithregardtoacongregationdisappearingorbeingsubjecttomerger(NCSdata)Disap-peared1Disap-peared2Disap-peared3Disap-peared4Disap-peared5Merged1Merged2Merged3Merged4Merged5PredictorsOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosIntercept0.22***0.28***3.36**2.88*8.99***0.10***0.07***0.08***0.06*0.102000–9999inh.0.690.52**0.700.650.56*0.26**0.31**0.33**0.32**0.42*10,000–99,999inh.0.830.51*0.640.580.53*0.07***0.11**0.12**0.12**0.35100,000+inh.0.950.560.780.620.570.601.081.101.343.73*Yearoffoundation1.70***1.29*1.151.120.63**0.67**0.70*0.86Numberregularparticipants0.65***0.72**0.61***0.700.720.73Numbertotalmembers0.82*0.890.871.281.240.88Renting1.311.670.26*1.24Owner0.680.760.430.53Meetingspace0.50***0.53**0.42*0.38*Conflict1.080.850.820.71Income0.980.971.141.00Catholics0.31*0.64ReformedProtestant0.30*12.39*EvangelicalCharismatic1.7447.26**EvangelicalConservative0.790.00OrthodoxChristian2.470.00(Continued) 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License16 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONTable3:(Continued)Disap-peared1Disap-peared2Disap-peared3Disap-peared4Disap-peared5Merged1Merged2Merged3Merged4Merged5PredictorsOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOddsRatiosOtherChristian0.680.00Jewish2.09*3.25Muslim0.805.20Buddhist0.440.00Hindu/Sikh0.590.00Others0.540.00Observations1022102210221022102210221022102210221022R2Tjur0.0030.0220.0900.1150.1650.0270.0420.0460.0710.137Deviance873.829853.747792.621771.634724.544310.460301.258297.384285.455246.097AIC881.829863.747806.621795.634770.544318.460311.258311.384309.455292.097∗ p<.05;∗∗p<.01;∗∗∗ p<.001. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 17In the three models for disappeared, we find that the immigrant variable has no significanteffect. In the Disappeared 2 model, the odds of congregations disappearing are 1.5 times higherin an urban context. However, controlling for religious tradition reduces this odds ratio againto below 1 (0.672). We have just noted that non-Christian traditions generally have a very highprobability of creating new congregations. At the same time, non-Christian communities are alsothe most likely to disappear. For example, the odds of appearing for Hindus are 4.6 times higherthan for the reference category, but their odds of disappearing are simultaneously 4.3 times higher.We find here the turnover phenomenon discussed above.In the three models depicting mergers, we find again that it is mostly Established Christians,Reformed Protestants in particular, that merge. Once religious tradition is controlled for (Merged3 model), a new insight emerges with the odds ratios for the different levels of community size.The odds of a congregation in a very large city merging are almost twice as high as those fora congregation in the reference category (0–1000 inhabitants). Conversely, the odds of con-gregations located in “intermediate” communities (2000–9999 and 10,000–99,999 inhabitants)merging are about half as high as those in the reference category.Table 3 uses NCS data and displays models explaining disappearances and mergers withseveral additional explanatory variables.10 Our findings from the census data are grosso modoconfirmed, but the NCS perspective allows additional insights. Disappearances are significantlymore likely if a congregation is small (measured by the number of regular participants), and if ithas no weekday meeting place (meeting space variable). Although it is not very surprising thata large pool of active participants protects the congregation from closure, it is more interestingthat meeting space has a very strong and significant effect on the congregation’s survival. In theDisappeared 4 and Disappeared 5 models, congregations that have a weekday meeting spacepresent roughly half the odds ratio of closure, signifying a considerably lower probability ofdissolution compared to thosewithout such a space.We can hypothesize thatmeeting spaces allowparticipants to obtain goods such as social connections and “informal social support” (Ellison1995), which in turn encourage them to keep the congregation alive.In line with previous studies (Dougherty, Maier, and Lugt 2008), models Disappeared 2 andDisappeared 3 show a negative correlation between the age of a congregation and its propensity todisappear (Year of foundation variable). However, in models 4 and 5, this correlation is no longersignificant. The total number of members, ownership of the building, conflict in the previousfew years, and income are generally not significant in predicting disappearance. In other words,resources have a mixed impact on diversity dynamics, with only the availability of a weekdaymeeting space and the number of regular participants being strong predictors of disappearance.As for mergers, the main finding is again that these are basically a phenomenon found amongReformed Protestants, and to a lesser extent among Catholics. Interestingly, the size of the con-gregation in terms of the number both of regular participants and members does not seem to berelated to the likelihood of a merger (Merged 3–5 models).ConclusionThe goal of our article has been to look at how religious diversity in Switzerland changedbetween 2008 and 2020, and to explore the mechanisms underlying these changes. To do so, wecombined data from two congregations censuses conducted in 2008 and 2020with a sample-basedNCS from 2008. Our research has led to several new findings: (1) Between 2008 and 2020, therewas no increase in religious diversity at an aggregate level in Switzerland. This finding remains10To verify the existence of collinearity problems in these models and these presented in Table 2, we submitted eachmodel with more than one predictor to a Generalized Variance Inflation Factor (GVIF) test. The resulting GVIF valuesfor predictors were found to be well below 10, an often-used threshold for collinearity detecting (Aheto 2019). 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License18 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONrobust when changing the measurement precision of religious tradition and controlling for urban-ism or language region. The finding that religious diversity remains stable over a period of morethan 10 years contrasts with many studies that report that religious diversity is increasing, espe-cially in urban contexts, which has sometimes been labeled “super-diversity” (Vertovec 2007).11(2) Although religious diversity is very stable at an aggregate level, the percentage of new, persist-ing, and defunct congregations varies strongly between religious denominations at an underlyinglevel. (3) Rural areas lose congregations mainly because established Christian groups merge theirrural parishes. Urban areas have a high turnover of congregations mainly due to the activities ofnonestablished Christians and non-Christian groups. Urban areas lose fewer congregations thanrural areas. (4) Overall, the proportion of immigrant communities has increased. Immigrationis the major reason why there is much appeared Orthodox and Muslim congregations. It is alsowidely significant for other traditions such as charismatic Evangelicals. However, not all pocketsof growth can be explained by migratory flows: only a small part of the new Buddhist, Hinduand Sikh congregations were the result of immigration during the time of observation. (5) Thedisappearance of congregations belonging to established denominations only occurs as a resultof mergers, while nonestablished groups hardly ever merge. Reformed Protestants and Catholicsalso have different strategies when it comes to merging congregations. While the former simplymerge, the latter often use a compromise strategy that sees unmerged parishes being combinedinto “pastoral units” that nonetheless function very much like merged parishes. (6) We find thatthe likelihood that a congregation will disappear increases especially when it lacks regular partic-ipants and, interestingly, a meeting space. Other resources-related features of the congregation,such as income or building ownership were however not significant determinants in our mod-els. Whether these findings can be generalized for other countries in western Europe is an openquestion, but we suspect that many can for countries such as Germany, France, and Italy, sinceindividual religiosity has evolved in these countries in a very similar way to how it has evolvedin Switzerland, and a remarkably similar picture emerges of these countries in various studies ofcongregations (Monnot and Stolz 2018a).From a distance, the declining number of congregations seem to suggest further that Swisssociety is subject to secularization. Although the Swiss population has grown significantly, thenumber of religious congregations has declined. This is the case across the country as a whole,but especially in rural areas, where the established churches have merged and continue to mergetheir parishes. Even in urban areas, the overall number of congregations is on the decline, withglobalization and the influx of new religions being unable (at least in the area and the periodobserved here) to compensate for the effects of secularization at a congregational level.There are obvious limitations to our study. First, we only compare our indicators for twotime points, and it is still open whether the changes found point to longer lasting trends. Futurewaves will give us answers here. Second, we have looked at diversity only at the mesolevel oflocal religious groups. Microlevel indicators of aggregate religious diversity are also important tograsp religious diversity.12 Third, there are inherent limitations to our data, since certain religious11We acknowledge that the concept developed by Vertovec (2007) was not only intended to mean “a lot of diversity,” butalso included various other elements. As Meissner (2015) has shown, however, many researchers used the term later inthe sense only of “a lot of diversity”. To be critical, though, we should also say that Vertovec neither clearly defined noroperationalized super-diversity (Meissner and Vertovec 2015; Vertovec 2007). Since Vertovec provided no guidance onhow to distinguish a situation of diversity from one of super-diversity, it is in our view still debatable as to whether weneed the new term at all.12By way of example, Liedhegener and Odermatt measured the evolution of HHI between 2000 and 2010 for 44 Europeancountries, based on individual affiliation. They do point to a “modest” (2018:42) religious pluralization. They found thatthe county-level median of 1 – HHI increases (from 0.38 to 0.43). However, they also note that most European countriesremain stable, that only five to seven countries (almost all Northern European) see their diversity increase substantially, 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 19traditions have few congregations, which often makes statistical evaluations difficult. Fourth, wehave ignored the issue of how the state influences religious diversity (Becci 2015).Apart from providing substantive findings, our study also allows us to formulate two furthergeneral insights. The first insight, which is obvious but important, is that longitudinal data are nec-essary to judge whether changes in diversity are significant. A one-shot study cannot ascertain anincrease in diversity, and nor can a retrospective study, since neither can measure the number ofdefunct congregations. The second insight is that, if we aim to interpret the changes undergoneor the stability in aggregate religious diversity, then it is important to address the mechanismsunderlying these processes. We need to record the number of new, persisting, and defunct con-gregations in different religious traditions, categorize the form that each takes, and relate themto determinants. Only by proceeding in such a manner can we reveal many important features ofthe process of religious change.DATA AVAILABILITY STATEMENTThe underlying data as well as the R scripts leading to the results in this article are availablefrom the authors upon request and will be made available in the following data repositories: FORS(https://forscenter.ch) and ARDA (https://thearda.com).FundingThe authors base themselves on data gathered due to the Swiss National Science Foundation(SNSF). Jeremy Senn’s salary is also funded by the SNSF.ReferencesAheto, J. M. K. (2019). Predictive model and determinants of under-five child mortality: evidence from the 2014 Ghanademographic and health survey. BMC Public Health 19(1). https://doi.org/10.1186/s12889-019-6390-4Anderson, Shawna L., Jessica Hamar Martinez, Catherine Hoegeman, Gary Adler, and Mark Chaves. 2008. Dearly de-parted: How often do congregations close? Journal for the Scientific Study of Religion 47(2):321–28.Bäckström, Anders, Grace Davie, Ninna Edgardh, and Per Pettersson. 2016.Welfare and religion in 21st century Europe:Volume 1: Configuring the connections. London: Routledge.Baumann, Martin. 2002. Migrant settlement, religion, and phases of diaspora. Exemplified by Hindu traditions steppingon European shores.Migration. A European Journal of International Migration and Ethnic Relations (33):93–117.Baumann,Martin. 2012. Religionsgemeinschaften imWandel - Strukturen, Identitäten, interreligiöse Beziehungen. InRe-ligionen, Staat und Gesellschaft. Die Schweiz zwischen Säkularisierung und religiöser Vielfalt, edited by ChristophBochinger, pp. 21–76, Zürich: NZZ.Becci, Irene. 2015. Institutional resistance to religious diversity in prisons: Comparative reflections based on studies inEastern Germany, Italy and Switzerland. International Journal of Politics, Culture, and Society 28(1):5–19.Becci, Irene, Marian Burchardt, and Mariachiara Giorda. 2016. Religious super-diversity and spatial strategies in twoEuropean cities. Current sociology 65(1):73–91.Berchier, Rémy. 2012. Regroupements paroissiaux, nouvelles paroisses, unités pastorales en Suisse romande. LumenVitae 67(1): 73–78.Bouma, Gary D. 1997. Many religions, all Australian: Religious settlement, identity and cultural diversity. Kew: TheChristian Research Association.Chaves, Mark. 2004. Congregations in America. Cambridge: Harvard University Press.Chaves, Mark. 2018. Congregations in Europe and the United States: Surprising similarities and common priorities forfuture research. In Congregations in Europe, edited by Christophe Monnot and Jörg Stolz, pp. 209–19, Cham:Springer.Dougherty, Kevin D, Jared Maier, and Brian Vander Lugt. 2008. When the final bell tolls: Patterns of church closings intwo protestant denominations. Review of Religious Research 50(1):49–73.and that four countries even endure opposite trends. Switzerland is one of the countries in which the HHI “was leftunchanged” (2018:45). 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License20 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONEck, Diana L. 2001.A new religious America. How a “Christian country” has become the world’s most religiously diversenation. New York: HarperCollins.Ellison, Christopher G. 1995. Rational choice explanations of individual religious behavior: Notes on the problem ofsocial embeddedness. Journal for the Scientific Study of Religion 34(1):89–97.Finke, Roger, and Rodney Stark. 2005. The churching of America, 1776–2005: Winners and losers in our religious econ-omy. New Brunswick, NJ: Rutgers University Press.Fons Duocastella, Clara, Blanca Luque Capellas, and María Forteza Gonzales. 2012. El mapa de minories religioses deCatalunya. Revista Catalana de Sociologia 28:12–27.Friederich, Ueli. 2017. Rechtliche und praktische Fragen zum Zusammenschluss von Kirchgemeinden. SchweizerischeJahrbuch für Kirchenrecht 22:11–59.Giordan, Giuseppe. 2018. Be fruitful and multiply… fast! The spread of Orthodox churches in Italy. In Congregations inEurope, edited by Christophe Monnot and Jörg Stolz, pp. 53–70, Cham: Springer.Harrelson, Walter. 2005. Evangelism. In A Dictionary of Jewish–Christian relations, edited by Edward Kessler and NeilWenborn. 150. Cambridge: Cambridge University Press.Hutchinson, Mark, and John Wolffe. 2012. A short history of global evangelicalism. Cambridge: Cambridge UniversityPress.Iannaccone, Laurence R. 1995. Voodoo economics? Reviewing the rational choice approach to religion. Journal for theScientific Study of Religion 34(1):76–88.Knott, Kim. 2015. Researching local and national pluralism: Britain’s new religious landscape. In Religiöser pluralismus,edited by Martin Behloul Bauman and M. Samuel, pp. 45–68, Bielefeld: transcript Verlag.Körs, Anna. 2018. Mapping congregations in Germany: Organization, beliefs, activities, and relations - the case study ofHamburg. In Congregations in Europe, edited by Christophe Monnot and Jörg Stolz, pp. 117–38, Cham: Springer.Krech, Volkhard. 2009. What are the impacts of religious diversity? A review of the methodological considerations andempirical findings of a research project on religious pluralisation in North Rhine-Westphalia, Germany. Religion39(2):132–46.Liedhegener, Antonius, and Anastas Odermatt. 2018. Religious affiliation as a baseline for religious diversity in contem-porary Europe. Making sense of numbers, wordings, and cultural meanings. Luzern: University of Lucerne.Martikainen, Tuomas. 2004. Immigrant religions in local society. Historical and contemporary perspectives in the city ofTurku. Åbo: Åbo Akademis förlag.Martinez-Arino, Julia. 2018. The evolution of religious diversity: Mapping religious minorities in Barcelona. In Congre-gations in Europe, edited by Christophe Monnot and Jörg Stolz, pp. 177–94, Cham: Springer.May, Matthew. 2018. The emergence and decline of Southern Baptist congregations in a southeastern county, 1784–2011:An ecological analysis. Review of Religious Research 60(4):455–75.Meissner, Fran. 2015. Migration in migration-related diversity? The nexus between superdiversity and migration studies.Ethnic and Racial Studies 38(4):556–67.Meissner, Fran, and Steven Vertovec. 2015. Comparing super-diversity. Ethnic and Racial Studies 38(4):541–55.Monnot, Christophe. 2013. Croire ensemble. Analyse institutionnelle du paysage religieux en Suisse. Zürich: Seismo.Monnot, Christophe. 2016. Visibility or invisibility: The dilemma of the muslim associations of Switzerland. Journal ofreligion in Europe 9(1):44–65.Monnot, Christophe, and Jörg Stolz. 2014. The diversity of religious diversity. Using census and NCS methodology inorder tomap and assess the religious diversity of a whole country. InReligious pluralism. Framing religious diversityin the contemporary world, edited by Giuseppe Giordan and Enzo Pace, pp. 73–91, Cham: Springer.Monnot, Christophe, and Jörg Stolz, eds. 2018a. Congregations in Europe. Cham: Springer.Monnot, Christophe, and Jörg Stolz. 2018b. How do you recognize a “congregation”? Definition and operationalizationstrategies of the Swiss congregation census. In Congregations in Europe, edited by Christophe Monnot and JörgStolz, pp. 15–31, Cham: Springer.Monnot, Christophe, and Boris Wernli. 2023. Moving away from religion: Age, cohort, or period effect? Evidence froma longitudinal survey in Switzerland. Religions 14(4):493.Murphree, Adam. 2018. The spiritual bottom line faith, strategy, and evangelical church planting. New York: New YorkUniversity.Olson, Daniel VA, Joey Marshall, Jong Hyun Jung, and David Voas. 2020. Sacred canopies or religious markets? Theeffect of county-level religious diversity on later changes in religious involvement. Journal for the Scientific Studyof Religion 59(2):227–46.Pace, Enzo. 2018. Religious congregations in Italy: Mapping the new pluralism. In Congregations in Europe, edited byChristophe Monnot and Jörg Stolz, pp. 139–58, Cham: Springer.Qvortrup Fibiger, M. C. (2009). The Danish pluralism project. Religion 39(2):169–75. https://doi.org/10.1016/j.religion.2009.01.004Rebenstorf, Hilke. 2018. Congregational development and organization-mindedness: Results from the German churchbarometer. In Congregations in Europe, edited by Christophe Monnot and Jörg Stolz, pp. 71–89, Cham: Springer. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 21Sahli, Markus, and Matthias D. Wüthrich. 2007. Wohnung Gottes oder Zweckgebäude? Ein beitrag zur Frage derKirchenumnutzung aus evangelischer Perspektive. In SEK Impuls. Bern: Schweizerischer Evangelischer Kirchen-bund SEK. https://edoc.unibas.ch/15173/Stark, Rodney, and Roger Finke. 2000. A theoretical model of religious economies. In Acts of faith: Explaining the humanside of religion, edited by Rodney Stark and Roger Finke, pp. 193–217, Berkeley: University of California Press.Steiner, Reto, and Marcel Notter. 2021. Verbreitung von Kirchgemeindefusionen in der Schweiz. VM Verwaltung &Management 27(2):51–58.Stolz, Jörg, and Mark Chaves. 2017. Does disestablishment lead to religious vitality? The case of Switzerland. BritishJournal of Sociology 2(69):412–35.Stolz, Jörg, and Jeremy Senn. 2022. Generationen abnehmenden Glaubens: Religion und Säkularisierung in der Schweiz1930–2020. In Religionstrends in der Schweiz: Religion, Spiritualität und Säkularität im gesellschaftlichen Wandel,edited by Jörg Stolz, Arnd Bünker, Antonius Liedhegener, Eva Baumann-Neuhaus, Irene Becci, Zhargalma Dan-darova Robert, Jeremy Senn, Pascal Tanner, Oliver Wäckerlig, and UrsWinter-Pfändler, 7–31. Wiesbaden: SpringerVS.Tomberlin, Jim, and Warren Bird. 2012. Better together: Making church mergers work. San Francisco: Jossey-Bass.Village, A., and L. J. Francis. (2021). Churches and faith: Attitude towards church buildings during the 2020 covid-19 lockdown among churchgoers in England. Ecclesial Practices 8(2):216–32. https://doi.org/10.1163/22144471-bja10025Vejrup Nielsen, Marie. 2018. Religious change and continuity in a Danish town: Results from a mapping project. InCongregations in Europe, edited by Christophe Monnot and Jörg Stolz, pp. 159–78, Cham: Springer.Vertovec, Steven. 2007. Super-diversity and its implications. Ethnic and Racial Studies 30(6):1024–54.Voas, David, Daniel V.A. Olson, and Alasdair Crockett. 2002. Religious pluralism and participation: Why previous re-search is wrong. American Sociological Review 67(2):212–30.AppendixA1. Absolute Numbers of Defunct, Persisting, New, and Merged Congregations (CensusData)Table A1Table A1: Numbers of defunct, persisting, new, and merged congregation in 2008 and 2020 ac-cording to religious tradition (census data)Tradition Year 2008 Year 2020 Defunct Persisting New MergedCatholic 1960 1818 149 1811 7 149Reformed Protestant 1073 928 162 911 17 162Evangelical (Classical) 561 540 88 473 67 7Evangelical (Charismatic) 720 725 236 484 241 3Evangelical (Conservative) 195 128 68 127 1 0Orthodox Christian 70 118 6 64 54 0Other Christians 775 683 207 568 115 3Jewish 34 29 6 28 1 1Muslim 361 363 84 277 86 0Buddhist 141 161 45 96 65 0Hinduists/Sikhs 185 174 75 110 64 1Others 266 216 98 168 48 2Total 6341 5883 1224 5117 766 328 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License22 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONA2. Coding of Religious Tradition VariablesTable A2Table A2: Coding of religious tradition variablesTrad CodingsTrad3: Established Christian, non-established Christian, non-Christian.Trad6: Christian, Jewish, Muslim, Buddhist, Hindu/Sikh, OtherTrad12: (Roman) Catholic, Reformed Protestant, Evangelical (Classical), Evangelical(Charismatic), Evangelical (Conservative), Orthodox Christian, OtherChristian, Jewish, Muslim, Buddhist, Hindu/Sikh, Others.Trad16: (Roman) Catholic, Christ Catholic, Reformed Protestant, Evangelical(Classical), Evangelical (Charismatic), Evangelical (Conservative), OrthodoxChristian, Messianic (e.g., Jehovah’s Witnesses, Latter Day Saints, 7th DayAdventists), Neo-Apostolic, Other Christian, Jewish, Muslim, Buddhist,Hindu/Sikh, Bahaï, Others.Trad35: Roman Catholic, Christ Catholic, Reformed Protestant, Neo-Pietist,Independent Pietist, Anabaptist, Baptist, Methodist, Quaker, Salvation Army,Darbyst, Pentecostal and other charismatic, Other evangelicals, Lutherans,Anglicans etc., Neo-apostolic, Messianic (Jehovah’s Witnesses, LSD etc.),Christian Orthodox, Christian Science, Other Christian, Jewish, Muslim,Sufi, Alevites, Bahaï, Buddhist, Hindu, Neo-pagan, Sikh and Sant Mat,Neo-revelation, Theosophy, Spiritist and mediumnic groups, Scientology,Rosicrucian, UFO movements, Others. 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseDIVERSITY DYNAMICS 23A3. Diversity Scores with Religious Tradition Variables of Different Specificity inSwitzerland, Community Size, and Swiss Language Regions (Census)Table A3Table A3: Diversity scores according to different religious classificationsYear Area Trad6 Trad12 Trad16 Trad352008 Switzerland 0.28 0.83 0.84 0.862020 Switzerland 0.29 0.83 0.84 0.862008 0–1999 inhabitants municipalities 0.09 0.63 0.63 0.632020 0–1999 inhabitants municipalities 0.08 0.62 0.62 0.622008 2000–9999 inhabitants municipalities 0.15 0.77 0.78 0.792020 2000–9999 inhabitants municipalities 0.17 0.78 0.78 0.792008 10,000–99,999 inhabitants municipalities 0.37 0.86 0.88 0.912020 10,000–99,999 inhabitants municipalities 0.36 0.86 0.88 0.912008 100,000+ inhabitants municipalities 0.5 0.88 0.9 0.932020 100,000+ inhabitants municipalities 0.5 0.88 0.9 0.932008 German-speaking 0.29 0.84 0.86 0.872020 German-speaking 0.3 0.84 0.85 0.872008 French-speaking 0.27 0.83 0.83 0.852020 French-speaking 0.3 0.83 0.84 0.862008 Italian-speaking 0.19 0.54 0.55 0.552020 Italian-speaking 0.16 0.57 0.57 0.57 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License24 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGIONA4. Number and Proportion of Immigrant Congregations by Religious Tradition and Year(Census)Table A4Table A4: Number and proportion of immigrant congregations in 2008 and 2020 according toreligious traditionYear Trad12 Domestic Immigration NA2008 Catholic 1712(87.3%)246 (12.6%) 2 (0.1%)2020 Catholic 1583(87.1%)233 (12.8%) 2 (0.1%)2008 Reformed Protestant 1064(99.2%)7 (0.7%) 2 (0.2%)2020 Reformed Protestant 921 (99.2%) 6 (0.6%) 1 (0.1%)2008 Evangelical (Classical) 533 (95%) 15 (2.7%) 13 (2.3%)2020 Evangelical (Classical) 506 (93.7%) 24 (4.4%) 10 (1.9%)2008 Evangelical (Charismatic) 375 (52.1%) 270 (37.5%) 75 (10.4%)2020 Evangelical (Charismatic) 407 (56.1%) 282 (38.9%) 36 (5%)2008 Evangelical(Conservative)191 (97.9%) 0 (0%) 4 (2.1%)2020 Evangelical(Conservative)127 (99.2%) 0 (0%) 1 (0.8%)2008 Orthodox Christian 4 (5.7%) 65 (92.9%) 1 (1.4%)2020 Orthodox Christian 4 (3.4%) 113 (95.8%) 1 (0.8%)2008 Other Christians 593 (76.5%) 159 (20.5%) 23 (3%)2020 Other Christians 490 (71.7%) 183 (26.8%) 10 (1.5%)2008 Jewish 18 (52.9%) 12 (35.3%) 4 (11.8%)2020 Jewish 15 (51.7%) 12 (41.4%) 2 (6.9%)2008 Muslim 32 (8.9%) 298 (82.5%) 31 (8.6%)2020 Muslim 42 (11.6%) 296 (81.5%) 25 (6.9%)2008 Buddhist 92 (65.2%) 11 (7.8%) 38 (27%)2020 Buddhist 119 (73.9%) 14 (8.7%) 28 (17.4%)2008 Hinduists/Sikhs 138 (74.6%) 28 (15.1%) 19 (10.3%)2020 Hinduists/Sikhs 120 (69%) 38 (21.8%) 16 (9.2%)2008 Others 224 (84.2%) 15 (5.6%) 27 (10.2%)2020 Others 172 (79.6%) 26 (12%) 18 (8.3%) 14685906, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jssr.12907, Wiley Online Library on [15/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License |
|---|