The world is now witnessing democratic backsliding during the so-called ‘autocratization wave’ (Lührmann and Lindberg Reference Lührmann and Lindberg2019). Since 2006, there has been a gradual decline of democracy (Diamond Reference Diamond2015, Reference Diamond2021, Reference Diamond2022) and recent reports on the state of democracy indicate that the democratic advances of the third wave have now been eradicated. According to the ‘Autocratization Changing Nature’ report issued by Varieties of Democracy (V-Dem) (Boese et al. Reference Boese2022: 6), the number of closed autocracies is on the rise, covering 26% of the world population, while electoral autocracy is the most common regime type, covering 44% of the world's population. Consequently, scholars have diverted their attention from democratization studies to autocratization studies, which tackle democratic backsliding in democratic (Bermeo Reference Bermeo2016; Ding and Slater Reference Ding and Slater2021; Erdmann and Kneuer Reference Erdmann, Kneuer, Erdmann and Kneuer2011; Lührmann and Lindberg Reference Lührmann and Lindberg2019), hybrid (Dresden and Howard Reference Dresden and Howard2016) and authoritarian (Sinkkonen Reference Sinkkonen2021) regimes.
Despite the prominent role that religion was thought to play in the third wave of democratization (Huntington Reference Huntington1991; Toft et al. Reference Toft, Philpott and Shah2011), religion has so far not occupied a prominent role in autocratization studies. Several studies have addressed the role of different actors in supporting or resisting this autocratic wave (e.g. Lotfy Reference Lotfy2022; Tomini et al. Reference Tomini, Gibril and Bochev2023), but religious actors were not part of their discussion or were simply treated as part of civil society (e.g. Rakner Reference Rakner2021; Sjögren Reference Sjögren2022). Given the importance of religious actors in democratization, at least historically, it is worth considering how their attitudes and actions contribute to current trends in autocratization.
Following David Waldner and Ellen Lust (Reference Waldner and Lust2018), backsliding is the decline in the qualities of democracy that is witnessed within any type of regime, whether it is a democratic or an autocratic regime. Consequently, I studied 52 cases of dominant religious institutions (those followed by the largest share of the population) in democratic, hybrid and authoritarian regimes and examined how the religious institutions responded to democratic backsliding after the third wave of democratization, from the early 1990s until 2021.Footnote 1 This was done by collecting religious institutions’ statements and monitoring their media mentions. Following Nancy Bermeo (Reference Bermeo2016: 8–13), I chose to focus on acts of executive aggrandizement (specifically amendments to presidential term limits) and military coups, which are strategies related to regime survival.Footnote 2 Therefore, if a religious institution supports the survival of the regime, it will support presidential term-limit amendments (removing or extending term limits) and/or object to military coup attempts.
Of the dominant religious institutions studied, I found that nearly 65% (34) of them opposed regime survival in the post-third wave period, while nearly 35% (18) did not. This begs the question, why do some dominant religious institutions oppose regime survival while others do not?
I argue that the opposition to regime survival by dominant religious institutions is a strategic response to the degree of their religious monopoly and the dominant religious groups’ historical commitment to the third wave of democratization – that is, whether they supported the third wave or not – which I refer to as their historical pro-democratizing role. This study adds to the literature by offering a way of studying the impact of religious producers’ religious and social goods on their supply of political stances and commitments. Moreover, this article contributes to the literature by demonstrating that religious institutions’ political offerings can be constrained by their historical commitments – that is, their role in the third wave of democratization.
Following much of the political economy literature, I focus on the role played by the religious institution followed by the greatest number of people in the nation.Footnote 3 I include a variety of dominant religious institutions: 44% of the data (23 religious institutions) are Catholic, 37% (19 religious institutions) are Islamic, 16% (six religious institutions) are Protestant, and the remainder (four religious institutions) are Eastern Orthodox and Buddhist. Thus, the article examines, using logistic regression, the effect of religious monopoly and the historical pro-democratizing role of 52 dominant religious institutions from different regions that have faced presidential term-limit amendments and/or military coups in the post-third wave period.
The article proceeds as follows: the first section briefly reviews the literature on religion and religious actors’ role in promoting or hindering democratic transitions. The second section tackles the theoretical argument and the first hypothesis of how religious competition affects the market for legitimacy. The third section discusses the second theoretical argument and the second hypothesis of how a historical pro-democratizing role for the religious institution affects a regime's legitimacy. The fourth section lays out the data and the methodological tools used in this article. The fifth section maps religious institutions’ stances towards autocratic acts according to the data used in this study. The sixth section tests the hypotheses in three different models and discusses the statistical and substantive significance of the results before concluding in a final section.
Religion in transitions
Several studies have tried to explain religious actors’ attitudes towards the third wave of democratization, with some explanations focused on religious doctrine. These studies claim that certain religions, such as Islam and Christian Orthodoxy, do not provide valid grounds for the development of democratic institutions (Fukuyama Reference Fukuyama1992; Huntington Reference Huntington1996; Kedourie Reference Kedourie1992). However, a group of studies disagreed with these theses (Anderson Reference Anderson2004; Ayubi Reference Ayubi1991; Filali-Ansary Reference Filali-Ansary1999), arguing that religion should not be treated as homogeneous since each religion or sect hosts a variety of schools of interpretation (Stepan Reference Stepan and Stepan2001).
Some also addressed changes in religious doctrines, arguing that one of the important changes that made the third democratic wave possible happened within the Catholic Church, when it adopted pro-democratic and pro-social justice stances (Huntington Reference Huntington1991; Mantilla Reference Mantilla2010; Thiede and Carnes Reference Thiede and Carnes2018; Troy Reference Troy2009). Moreover, some studied the state–religion relationship, arguing that the condition most affecting the role of religious actors in all three waves of democratic transitions is the actual autonomy they enjoyed (Jamal Reference Jamal2009; Kunkler and Leininger Reference Kunkler and Leininger2009; Philpott Reference Philpott2004, Reference Philpott2007; Stepan Reference Stepan2000).
Other scholars took a religious competition/political economy approach, but they differed in the type of religious markets that would drive political stances. Some of these studies only analysed the effects of religious competition in Latin American countries (Froehle Reference Froehle1994; Gill Reference Gill1994, Reference Gill1998; Smith Reference Smith2016; Trejo Reference Trejo2009), arguing that Catholic churches took liberal stances because of religious competition with Protestants – to prevent conversions to the competing religion. A few have developed these dynamics in democratic countries (Djupe and Neiheisel Reference Djupe and Neiheisel2019; Monsma and Soper Reference Monsma and Soper2009; Rosenberg and Smith Reference Rosenberg and Smith2021), where they found that the political goods that the churches provide are directed to attract new members and retain the current ones. Karrie Koesel (Reference Koesel2015) assessed the religious economy approach in Russia and China, where her study revealed similar political dynamics under heavy market constraints. Meanwhile, Ryan J. Rebe (Reference Rebe2012) focused on religious pluralism in democratic and non-democratic countries worldwide, finding that societies that develop institutions promoting religious pluralism inhibit dominant religions from forming alliances with the political elite.
Therefore, while it is possible that religious actors’ political appeal to the masses could be witnessed in competitive and monopolist markets, the majority of research on religious markets finds that liberal stances are more likely in competitive markets. The scholarship then shifted to focus on the role of religious actors in the democratization period itself and how this process could build religion-friendly institutions that would protect space for religion and religious actors to form political parties and compete in elections; this is what Michael Driessen (Reference Driessen2014a) called ‘religiously friendly democratization’.
But democratization and economic liberalization also have effects on the religious market, with research arguing that they lead to the proliferation of fewer institutionalized religious actors, often charismatic or reformist, who are seen as challengers to the established religions or as upcoming political forces. Many of these studies focused on the African region (e.g. Abbink Reference Abbink2014; Cooke and Downie Reference Cooke and Downie2015) and the Middle East (e.g. Albayrak Reference Albayrak2019; Cesari Reference Cesari2014; Driessen Reference Driessen2014a, Reference Driessen2014b; Gurses et al. Reference Gurses, Fox and Ozturk2023).
Existing research does not focus on the stances of the main institutions of these religious groups towards transitions to autocracy and/or deepening authoritarianism after the third wave of democratization and after the changes that happened in the religious market in the post-third wave period. Furthermore, the political economy of religious market approach has not addressed the relationship between the degree of religious competition and the historical commitment of the religious groups in supporting democratic transitions or authoritarianism. This article tries to fill this gap.
Religious competition, regulation and the market for legitimacy
Religious competition has been growing across the world. According to a 2014 Pew Research Center poll (Masci Reference Masci2014), tens of millions of Latin Americans who left the Roman Catholic Church have converted to Pentecostalism. These switches have also been embraced in Africa, where new charismatic Pentecostal churches have become prominent since the 1980s (Adeboye Reference Adeboye, Afolayan, Yacob-Haliso and Falola2018: 26–30). These religious upstarts have been a challenge to established religious institutions in these regions. In consequence, Protestants have been driving the Catholic Church to cater to the people and take more liberal stands, as Anthony Gill (Reference Gill1994, Reference Gill1998) asserted. Meanwhile, apart from Latin America and Africa, the communist and post-Soviet worlds have seen a resurgence of religion after the collapse of the Soviet Union as the ruling elite sought another ideology to rely on to fill the void of legitimacy (Koesel Reference Koesel2015: 212).
To understand the implications of this religious resurgence and competition, this study relies on the religious economy approach, which assumes that the religious market works like the commercial one, where it benefits from competition and is endangered if there is a monopoly by one religion and there is high state regulation of religion (Iannaccone Reference Iannaccone1991). Some scholars found that market competition between secular and religious ideologies affected the presence of religion in the public space (Fox Reference Fox2016, Reference Fox2019b; Kortmann Reference Kortmann2019), while others focused on the inter-religious competition in the religious market (Gill Reference Gill1994; Iannaccone Reference Iannaccone1991; Stark and Iannaccone Reference Stark and Iannaccone1994). In this article, the focus is on the latter, examining how competition between different religious denominations in the religious market affects support for the ruling regime.
Religious markets consist of current and potential members, current and potential religious producers, and product lines offered by one, a few or various religious producers (Stark and Iannaccone Reference Stark and Iannaccone1994: 232). Religious competition is the presence of multiple religious producers in a community and can range from low in the case of monopolies to high in the case of pluralism. When the state does not regulate religion in either of two possible ways – suppression, in which the state penalizes religious groups, or subsidy, in which the state rewards one or more selected religious groups – free religious competition exists (Finke Reference Finke and Young1997: 50). Both forms are state intervention, raising the costs of joining an alternative religion (Finke Reference Finke and Young1997: 52). Thus, high state regulation, which discriminates against minority religious groups, helps to maintain a low degree of religious pluralism (Fox and Tabory Reference Fox and Tabory2008), and when the government supports a certain religion, religious institutions become highly dependent on it and vulnerable to its control (Fox and Breslawski Reference Fox and Breslawski2023: 6).
Following supply-side theories, the existence of many religious alternatives does not guarantee religious competition unless there are fewer restrictions from public authorities. Similarly, the existence of one religion that manipulates the market of salvation does not guarantee a religious monopoly unless there are higher religious restrictions from public authorities. Thus, if more than 50% of the population is affiliated with the dominant religious institution and there is a high level of restriction from public authorities, this would guarantee a higher degree of monopoly in the religious market (Case 1). However, if the dominant religious institution represents less than 50% of the population and there are few restrictions from public authorities, this would guarantee a higher degree of competition in the religious market (Case 2).
Meanwhile, if more than 50% of the population is affiliated with the dominant religious institution and there are few religious restrictions imposed by public authorities, this would represent a threat to the dominant religious institution as it leaves the market open to competitors (Case 3). Similarly, if the dominant religious institution represents less than 50% of the population and there are high religious restrictions imposed by public authorities, this does not guarantee a competitive religious market, as the religious consumers will not be able to switch their religion. Therefore, the religious market will be divided into small non-competitive markets, where each religious denomination has its own segment of the market (Case 4). As such, Cases 1 and 4 represent a higher degree of religious monopoly, while Cases 2 and 3 represent a lower degree of religious monopoly and more competition.
In the religious market, religious producers offer both religious and social goods. Religious goods come in the form of expressions of faith, promises of otherworldly rewards, moral guidance and answers to the questions of life and death (Gill Reference Gill1994: 405). Relying on these intangible products alone to attract and retain loyal members is too risky for religious producers (Gill Reference Gill1994: 407). Therefore, religious producers are engaged in providing social goods related to community service projects, especially those that are related to the lower classes, for example, literacy campaigns, job training, agrarian projects, medical and financial assistance, and the establishment of schools and hospitals (Gill Reference Gill1994: 412–416).
These goods offered by religious producers can impinge on state legitimacy as well. There are intangible and tangible political goods that are produced to legitimize the government and opposition movements. The intangible political goods are in the form of lending legitimacy and linking religious beliefs and symbols to the government or the opposition (Fox Reference Fox2018: 75), while the tangible political goods come in the form of financial and organizational resources, direct access to the media, and international connections (Fox Reference Fox2018: 74). These tangible and intangible goods are used for political mobilizations that could serve the interests of the government or the opposition. For example, the direct engagement by religious actors in elections, known as ‘religious politicking’, is when religious institutions engage with civil society groups to observe the elections and campaign for a candidate during religious services (Rosenberg and Smith Reference Rosenberg and Smith2021: 737–738).
Therefore, religious institutions, with their dense organizational network and symbols, can provide political actors with a base for political mobilization (Froehle Reference Froehle1994: 160). This political mobilization could be used to legitimize the regime's attempt to remain in power or could be used by the opposition and civil society actors to legitimize the removal of the regime. Religious institutions can support regime survival when they legitimize the attempt to amend the constitution to prolong the ruler's tenure, and they can oppose regime survival, for example, when they support military coups. Most of these current military coups are known as promissory coups (Bermeo Reference Bermeo2016) or civil society coups (Encarnación Reference Encarnación2002), where they are carried out to restore democracy or take place against corrupted regimes and are supported by the opposition and civil society actors – coup-friendly institutions (Eynde Reference Eynde2011).
Therefore, the political goods that the religious institutions provide are produced in complementarity with the religious and social goods to preserve their position in the religious market. In economics, ‘complement in production’ refers to two or more goods that are jointly produced from the same resource (McAlister and Lattin Reference McAlister and Lattin1983: 4). An increase in the price of one good will increase the supply of another good. When a monopolistic main religious institution in the religious market charges a high price for the religious and social goods it provides, the supply of the political goods it provides for the government increases. Thus, the monopolist religious institution will serve the interests of the incumbent regime by developing theological interpretations and practices that justify the need for the regime to remain in power (Trejo Reference Trejo2009: 325).
Nevertheless, the freedom of entry of one or more competitors in the religious market will reduce the price of enjoying the aforementioned products in the religious market; as a consequence, the main religious institution will have to reduce its prices as well (Cantoni et al. Reference Cantoni, Dittmar and Yuchtman2018: 2048), which means it reduces the costs of money, time, strict behavioural demands (Finke Reference Finke and Young1997: 53) and religious clothing – that is, lower membership requirements. Any reduction in the membership requirements lessens the effectiveness of religious goods production (Finke Reference Finke and Young1997: 54).
Alternatively, when the religious institution fears reducing the costs of membership in order not to lose its religious legitimacy, it starts to look for a substitute to save its position in the marketplace. The religious institution will compensate for its loss of religious and social goods by increasing its political goods by acting as a voice for the poor, dissidents, democracy and constitutional defenders. Here, the religious institution will serve theological reinterpretations for its own preferences that mostly adhere to widely popular opinions in order to retain and attract members (Trejo Reference Trejo2009: 325) – even if, for example, the widely popular opinions support a military coup. These theoretical arguments lead to the following hypothesis:
Hypothesis 1: As the religious monopoly decreases in the religious market, the dominant religious institution is more likely to object to regime survival.
It is worth noting that religious competition can occur in political regimes that were not necessarily fully democratic. Moreover, religious competition is an important ingredient for the development of democracy (Rebe Reference Rebe2012: 659). If the start of the religious competition mostly happened in a non-democratic environment, it can help nurture a future democratic state.
The historical pro-democratizing role and the market for legitimacy
The democratic reputation of the religious actors, whether they served the interests of authoritarian regimes or participated in the calls for democratization during the third wave, is also important to study for its effect on religious actors’ stances towards the current calls to remove the incumbent regimes. At issue is whether religious traditions are sticky or adaptable in response to incentives. For instance, the literature has not yet tackled whether religious actors benefited from their roles as democratic seekers or preservers of the status quo. In religious economies, the level of regulation and its nature – whether supportive or restrictive – are the sole factors that affect the incentives for religious producers (Finke Reference Finke and Young1997: 50). The historical commitment of the religious groups, which demonstrates the expenses they have incurred or the advantages they have gained from their responsibilities throughout the third wave, is something this article adds to the self-interest thesis of religious economies.
Monica Toft et al. (Reference Toft, Philpott and Shah2011: 86) studied the position of religious actors in the transition period of the third wave of democratization. They found religious actors who played supportive roles in the democratization process; those who played a leading, rather than a supportive, role in defending and promoting democratic governance; and religious actors who were indifferent but used this expanded freedom to increase their influence – ‘free riders’. There are also religious actors who have taken steps to slow, mitigate or impede the transition to democracy, either through their active support for an authoritarian regime or through their opposition to the establishment of democratic institutions – that is, their role was reactionary resistance.
Hence, religious groups who played a democratizing role, such as protesting against an authoritarian regime, cooperating with international actors to weaken an authoritarian regime, supporting opposition groups or mediating between political actors to facilitate democratic transitions (Toft et al. Reference Toft, Philpott and Shah2011: 94–95) are more likely to oppose leaders with autocratic-intent or to support military coups intended to restore democracy in the post-third wave period. However, religious groups are more likely to play an anti-democratic role when they lent former authoritarian regimes symbolic or material support or passively followed the lead of opposition actors during the third wave of democratization (Toft et al. Reference Toft, Philpott and Shah2011: 109). This theoretical argument leads to the following hypothesis:
Hypothesis 2: The dominant religious institution is more likely to object to regime survival if the dominant religious group supported democratization in the third wave.
I also examine whether the commitment to the historical pro-democratic stance is effective in any religious market or if it is only enforced and shaped in certain religious (monopoly or competitive) markets. Therefore, studying the interaction between the status of the religious market and the stability of the commitment is also important in order to understand the changing nature of the context since the third wave of democratization. I expect that free markets enable religious groups to maintain historical commitments.
Data and method
The analysis requires finding religious institutions that are faced with the decision whether or not to oppose the regime after the third wave – this determines the population. From there, I searched carefully for evidence of religious institutions taking stances. That process determined the observations studied in this article.Footnote 4 The time period for data collection is from the early 1990s, when the third democratic regression started (Lührmann and Lindberg Reference Lührmann and Lindberg2019), until 2021. According to the data in the term-limit literature, which tackled the phenomenon across regions (Lotfy Reference Lotfy2022; McKie Reference McKie2019), just over 70 presidents attempted to contravene term limits in the studied period. I managed to gather credible information about religious institutions’ stances in 38 attempts with an adequate distribution across regions.
Additionally, according to the Center for Systemic Peace's Coups d’État dataset (Marshall and Marshall Reference Marshall and Marshall2022), there were just over 20 successful attempted coups in the studied period. I managed to gather credible information about religious institutions’ stances in nine successful coups across regions but also added five short-lived coups that stimulated religious institutions’ interventions and popular protests, as happened in Venezuela in 2002 and Burkina Faso in 2015. (See Table A1 in the Supplementary Material for more details on the studied cases.)
Therefore, the current data include 52 dominant religious institutions from countries in Latin America (13 cases), the Middle East and Sub-Saharan Africa (28 cases), Asia-Pacific (5 cases) and post-Soviet regions (6 cases). These regions were chosen because they had long been subject to big-man or military rule. Nevertheless, the third wave of democratization came across these regions, and transitions to democracy (not necessarily full democracy) were instituted. However, through presidential term-limit amendments and coup d’état attempts, big-man rule and military interventions have been reintroduced in many of the countries in these regions. Religious institutions played a significant role in these regions during the third wave, and it is important to consider their role during the current democratic setbacks.
Dependent variable
The dependent variable is the stance of the dominant religious institutions towards autocratic practices related to regime survival. The dependent variable takes the value of 1 if the religious institution publicly opposed presidential term-limit amendments and/or supported a military coup – that is, opposed the incumbent regime's survival. The dependent variable takes the value of 0 if the religious institution did not oppose presidential term-limit amendments and/or did oppose the military coup – that is, did not oppose the incumbent regime's survival.Footnote 5 Thus, it is important to note that this was not measured by an absence of evidence, but only with direct evidence that they supported or did not oppose regime survival.Footnote 6
The data were gathered by collecting religious institutions’ statements and watching their media mentions in international and local news and on religious websites. The unit of analysis is the institutional level, which means that the focus is on the stances of the religious institutions in the national context. That includes the stances of the Episcopal Conference of Bishops, Catholic and Protestant churches, Islamic councils, grand mosques, Buddhist temples, and so on. For instance, the Catholic Bishops Conference of Burundi (2015) is coded as standing in opposition to the regime's survival as it stated in its second pastoral letter ahead of the 2015 elections that ‘The president elected by the people of Burundi must not exceed the two terms of five years stated in the constitution.’
If the individual religious leaders and the religious institutions have different stances, this study codes the stance of the religious institution. I do not assert that religious institutions are monolithic entities with religious leaders all acting similarly within the same institution. However, the bargaining power of the stance of the religious institution is typically more influential than the stance of a few defecting religious leaders. For instance, the Bolivian cardinal, Toribio Ticona, endorsed the re-election of Evo Morales, but a statement released by the Episcopal Conference indicated that this was a personal opinion that did not represent the Church, which rejected the 2017 court ruling that allowed Morales to run for re-election (Martín Reference Martín2019). Therefore, this study codes the Episcopal Conference's stance.
Independent variables
The two main independent explanatory variables are religious monopoly and the historical pro-democratizing role in the third wave. Religious monopoly is defined as the absence or scarcity of religious alternatives, as well as the inability to switch from one religion to another in the presence of religious alternatives. It is measured by constructing an interaction term that consists of the percentage of the national population identified with the dominant religious institution (its scale ranges from 0 to 1) and the degree of religious regulation (its scale ranges from 0 to 1).Footnote 7 Thus, the scale of the religious monopoly variable ranges from 0 to 1.
Data on the percentage of the dominant denomination come from the International Religious Freedom Reports created by the Office of International Religious Freedom in the US Department of State, which gathers information on religious demography from governmental and non-profit organization sources in several countries. Some of the data are also collected from the Association of Religion Data Archives (ARDA) National Profiles (Johnson and Grim Reference Johnson and Grim2022). Data on religious regulation are taken from V-Dem – the chosen variable is ‘freedom of religion’ (Coppedge et al. Reference Coppedge2022; Pemstein et al. Reference Pemstein2022). This variable was chosen specifically because it measures religious conversion by individuals or groups who have the right to choose a religion and proselytize and practise it in public or private, and whether there are instances of discrimination against individuals or groups due to their religion.Footnote 8
The second main independent variable is the historical pro-democratizing role in the third wave, in which the dominant religious institution takes the value of 1 if the dominant religious group had a pro-democratizing role in the third wave of democratization and 0 otherwise. These data come from Toft et al. (Reference Toft, Philpott and Shah2011).
In addition, control variables are included, such as the annual percentage growth rate of the GDP per capita (which is lagged one year before the autocratic attempt). According to previous studies (Buckley and Mantilla Reference Buckley and Mantilla2013; Fox Reference Fox2007), a higher GDP per capita is associated with greater state capacity to regulate religion. The data are taken from the World Bank. Government favouritism towards dominant religions (official status, religious education, financial support, etc.) is also included as a control and lagged by one year. These data are based on Government Preference Religion 2.0 (Brown Reference Brown2020).Footnote 9 The incumbent ruler's religion, whether it is the same as the dominant religious group (takes the value of 1) or not (takes the value of 0), is also included. I collected the data through different media outlets. In addition, I also control for the religious family of the dominant religious institution.Footnote 10 A variable that measures discrimination against religious minorities is not included due to the high collinearity with the religious monopoly variable (about 0.7 correlation) (see Figure A2 in the Supplementary Material).Footnote 11 Therefore, this implies that the religious monopoly variable includes a dimension of discrimination against religious minorities.
Statistics concerning the means, standard deviations and ranges of the dependent and independent variables are listed in Table 1. Logistic regression models are used to test the hypotheses of this study. This type of model is appropriate for the binary form of the dependent variable with cross-sectional cases. But before attempting to test the hypotheses, the next section maps the stances of the dominant religious institutions towards regime survival in order to have a clear understanding of the data.
Descriptive Statistics
Descriptive comparative analysis
This section aims to map religious institutions’ stances towards autocratic acts related to regime survival according to the data used in this study. This is done by showing how dominant religious institutions, those with the largest population, reacted to the term-limit amendments and military coups in the post-third wave period.
As Table 2 shows, nearly 61% of the term-limit amendment attempts witnessed opposition from the dominant religious institutions. Meanwhile, nearly 79% of those which witnessed military coups in their country did not oppose them. This constitutes strong evidence that most dominant religious institutions took stances in opposition to regimes with autocratic intent in the post-third wave era.
Cross-Tabulation of the Components of the Dependent Variable
This raises questions about the contradictions in religious institutions’ stances towards autocratic acts in the post-third wave period. This could be explained by the fact that these military coup attempts, as previously mentioned in the theoretical section, were known as civil society coups d’état or promissory coups, which were actually launched against regimes with autocratic intent. One of the well-known examples is the Catholic Church in Venezuela, which supported the short-lived military takeover in 2002 that deposed President Chavez on the grounds that he had misused his position of authority and undermined democratic institutions (Encarnación Reference Encarnación2002). Nevertheless, other dominant religious institutions that opposed military coups condemned the coups as illegal and unconstitutional, as did the Islamic institutions in Burkina Faso in 2015 (Burkina24 2015), and as did the Methodist Church in Fiji in 2006 (Newland Reference Newland, Fraenkel, Firth and Lal2009: 187).
This section also considers the debate in the literature regarding which type of religion is democracy-friendly (Anderson Reference Anderson2004; Ayubi Reference Ayubi1991; Filali-Ansary Reference Filali-Ansary1999; Fukuyama Reference Fukuyama1992; Huntington Reference Huntington1996; Kedourie Reference Kedourie1992; Stepan Reference Stepan and Stepan2001), but here I am not focusing on the religion per se but on the institutions that represent these religions. Table 3 compares the stances of different types of religious institutions, and it reveals that Catholic institutions have a higher probability of opposing regime survival than the other religious families: 95% of Catholic institutions that witnessed term-limit amendments in their countries opposed them, while the only three military coup attempts in the data that Catholic majorities witnessed were supported by their institutions.
Comparison of Different Types of Religious Institutions
One explanation for these differences between Catholic and other institutions could be their internal organization. Roman Catholic institutions are more hierarchal and centralized, strengthening their organizational capacity to speak with a single voice (Sarkissian Reference Sarkissian2006: 159–61). Roman Catholics are primarily governed by the Vatican, which is headed by the pope (Sarkissian Reference Sarkissian2006: 163). The pope has a substantive say when Roman Catholic majority countries are in crisis. For example, Paraguayan ex-president Horacio Cartes claimed in 2017 that he reversed his decision to run for re-election during a contentious term-limit amendment attempt in response to Pope Francis's call to end violence and seek solutions (Blair Reference Blair2017). As a result, the pope's stance is now more critical in the eyes of the world, calling for peace and preserving democracy. This pattern is in fact new and does not accord with Catholic behaviour during the third wave and before, as there was considerable variation in Catholic stances towards authoritarian regimes (Edmonds Reference Edmonds2013: 644–645; Philpott Reference Philpott2004: 41–43).
There is only one case where the Catholic Church did not oppose regime survival. This case is the Roman Catholic Church in Rwanda, which did not show opposition to the term-limit extension for President Paul Kagame in 2015. The 1994 genocide of ethnic Tutsi in Rwanda is considered to be one of the main explanations for the Catholic Church's stance. The Catholic Church was accused of being a killing field within which the Tutsis were slaughtered, and even some pastors and priests were involved in this slaughter (Longman Reference Longman2001: 163). Three bishops and 12 clergymen were then murdered by soldiers of the Rwanda Patriotic Front (RPF) in 1994 as a form of revenge (All Africa 2008). Consequently, tensions between the Catholic Church and the state were raised, and the Catholic Church has lost moral authority in the eyes of the Rwandan people; it could not then claim to be a constitutional defender or a proponent of democracy, especially since Kagame's term-limit amendment received widespread popular support (Lotfy Reference Lotfy2022: 583).
Apart from the Catholic Church's stances, Table 3 shows no clear pattern for the Protestant institutions’ stances as 50% of the data opposed regime survival. It is worth noting that in its early days, the Pentecostal Church avoided interfering in politics, allowing Catholic Churches to be the voice of the weak and poor in protesting the government's corruption and human rights violations (Onongha Reference Onongha, Afolayan, Yacob-Haliso and Falola2018: 375). However, as Pentecostals’ numbers and influence grew and evolved, they became more involved in local politics in order to retain their majority and growing influence (Burgess Reference Burgess2015). In regard to Islam, it seems that Islamic institutions are more flexible, as 83% of Islamic institutions witnessing term-limit amendments in their countries did not oppose them. Meanwhile, 86% of those which witnessed military coups in their countries supported them.
Therefore, I control for the effect of the different types of religious institutions when testing the hypotheses, and it is expected that Catholic institutions are more likely to oppose regime survival than the other types of religious institutions.
Testing the hypotheses: statistical analysis and substantive results
To test the hypotheses, three models are constructed (see Table 4).Footnote 12 The first model exclusively tests the main independent variables, which are the religious monopoly and historical pro-democratizing role variables; the second model adds government favouritism, GDP per capita (annual % growth rate) and the ruler's religion; the third model adds a nominal categorical variable which is the religious families – Islamic, ProtestantFootnote 13 and other religions – while Catholic is the reference group.
Three Logistic Models for Religious Institutions’ Opposition to Regime Survival
Note: Bold values represent statistically significant results.
The results in the first and second models in Table 4 support the two hypotheses. The first hypothesis predicts that an increase in the religious monopoly score decreases the probability of opposing regime survival, while the second hypothesis predicts that the probability of opposing the regime increases when the dominant religious group had a pro-democratizing role in the third wave of democratization.
As the odds ratio in Model 1 (Table 4) shows, for each additional score in the religious monopoly, the odds of the dominant religious institutions opposing the regime decreased significantly by a factor of 0.01 (p < 0.01), while all other variables remain constant. A standard deviation increase in the religious monopoly (about 0.23) decreases the odds of the religious institution opposing the regime by a factor of 0.38 (z = −2.689, p < 0.01) (see Table A6 in the Supplementary Material). All else being equal, dominant religious groups that had a pro-democratizing role in the third wave of democratization have odds of opposing the regime that are 7.21 times higher than those that did not have a pro-democratizing role (p < 0.01).
Moreover, in Model 2, for each additional score in the religious monopoly measure,Footnote 14 the odds of opposition to the regime by the dominant religious institutions will decrease by a factor of 0.01 (p = 0.018), holding all other variables constant. A standard deviation increase in the religious monopoly (about 0.23) decreases the odds of the religious institution opposing the regime by a factor of 0.38 (z = −2.360, p = 0.018) (see Table A7 in the Supplementary Material). All else being equal, religious groups that had a pro-democratizing role in the third wave of democratization have odds of opposing the regime that are 9.51 times higher than those that did not have a pro-democratizing role (p < 0.01). No other variables have significant effects. Figure 1 shows the substantive significance of Model 2's estimates, using 84% confidence intervals (CIs),Footnote 15 illustrating the predicted probabilities of opposing regime survival by the two main explanatory variables: the religious monopoly score and the historical pro-democratizing role, while holding other variables at their means.
The Substantive Significance of Model 2, Using 84% CIs
Figure 1 shows that for a dominant religious institution, when other variables are held at their means (GDP annual growth is 2.2%, government favouritism is 2.13, and the ruler's religion is the same as the dominant religious institution), having a historical pro-democratizing role increases the probability of opposing regime survival by about 0.3 (see Table A8 in the Supplementary Material) when the religious monopoly score is below 0.63. When the monopoly score is above 0.63, there is no significant difference between dominant religious institutions that had a pro-democratizing role and those that did not. As such, the historical pro-democratizing role seems to be particularly useful when religious monopoly is low or moderate. For both pro-democratizers and non-pro-democratizers, the probability of opposing regime survival decreases as the score of religious monopoly increases. This effect is significant for pro-democratizers, whose predicted probability of opposing the regime decreases by 0.62 going from a 0.07 to a 0.9 religious monopoly score (84% CI: −0.997, −0.241). This effect is also significant for non-pro-democratizers, whose predicted probability of opposing the regime decreases by 0.61 going from a 0.07 to a 0.9 religious monopoly score (84% CI: −0.826, −0.387).
In addition, when the interaction between the religious monopoly and the pro-democratizing role is included in Model 2 (Table 4), Figure 2 shows that the slope of the pro-democratizing role is positive and significantly different from 0 when the religious monopoly is low or moderate.
Simple Slope Analysis for the Interaction between Religious Monopoly and Historical Pro-Democratizing Role
The results in Model 3 (Table 4) show that the religious monopoly variable lost its significance when the categorical variable of the types of religious institutions were added. In addition, the historical pro-democratizing role still has positive significance, but the p-value is equal to 0.07 (90% CIs). Meanwhile, the Catholic institutions, the reference group, seem to have a strong significance as the odds of opposing regime survival are 0.04 lower if it is an Islamic or Protestant institution than if it is a Catholic institution. This result was expected since 96% of the Catholic institutions opposed regime survival, as previously shown in Table 3.
However, the results of Model 3 do not contradict the first or the second hypotheses. Figure 3 shows that Catholic and Protestant institutions that opposed regime survival are facing high religious competition.Footnote 16 However, the Protestants who did not oppose regime survival are also facing the same competition as the Protestants who opposed. The determinant factor here for why some of these Protestant institutions opposed regime survival while others did not, even though both of them are witnessing high religious competition, is their historical pro-democratizing role.
The Density of Religious Families’ Stances by Religious Monopoly Score
Figure A6 (in the Supplementary Material) shows that most of the Protestant institutions in the data with a historical pro-democratizing role are the ones who opposed regime survival. This confirms the substantive results that were previously shown in Figure 1, which illustrate that having a historical pro-democratizing role increases the probability of opposing regime survival when the religious monopoly score is low or moderate.
One outlier here in the data is the case of Nigeria, where the Pentecostal groups had a pro-democratizing role against the military regime of Muslim president Ibrahim Babangida, fearing competition with Islam. Thus, they played a critical role as a pro-democracy group in the late 1990s (Burgess Reference Burgess2015: 44). However, the Pentecostal Fellowship of Nigeria did not oppose Olusegun Obasanjo's third-term bid in 2006, as evidenced by its president, Pastor Ayo Oritsejafor, who declared that ‘a third term in office is not a crime’ despite the fact that there were other prominent Pentecostal leaders who opposed this third-term bid (All Africa 2006), along with the Christian Association of Nigeria (Washington Post 2006).
Apart from Catholic and Protestant institutions, Figure 3 shows at first glance that Islamic institutions may not be affected by the scores of the religious monopoly. A closer look at the density of the Islamic cases in Figure 4 shows more cases in highly competitive markets (with a religious monopoly score less than 0.7) opposing regime survival, while there are more cases in highly monopolized markets (with a religious monopoly score greater than 0.7) that did not oppose regime survival compared to those who did.
The Density of Islamic Institutions’ Stances by Religious Monopoly Score
Conclusion
While scholars have devoted significant attention to the role of religious institutions in democratization, less attention has been given to their role in autocratization. This article set out to test the impact of religious monopoly and the historical pro-democratizing role on religious institutions’ stances towards autocratic practices related to regime survival in the post-third wave period. Religious economy approaches suggest that religious institutions are flexible to offer whatever is of interest to the marketplace, but here the role they played in the third wave of democratization suggests a stable commitment. The findings demonstrate that the historical pro-democratizing role commitment reinforces the opposing stances of religious institutions when there is high or moderate religious competition.
Moreover, this article sheds light on the stances of different types of religious institutions. It is shown that there is a mostly united opposition stance taken today by Catholic majorities across regions, contrary to the Islamic and Protestant majority countries, which vary in their stances towards regime survival in the post-third wave period. Therefore, it seems that the mostly united opposition stance taken today by Catholic institutions across regions is a lesson learned from the past. In order to maintain their position as the majority religion in the country, their role is to stand with the poor and preserve democracy. The religious freedom that followed the third wave of democratization also opened the space for religious competition in some of the Muslim-majority countries and the newly Protestant-majority countries. However, having a historical pro-democratizing role strengthens the probability of opposing the survival of the authoritarian-minded regimes. This suggests that religious competition urges religious institutions to maintain a stronger brand that remains consistent with prior practice.
Future research can re-examine these arguments by increasing the sample size to include more cases or focusing on limited cases qualitatively. Future research could also tackle the stances of minority religious institutions towards regime survival to assess if their stances complement or contradict the plurality of religions. Future research can also address the differences in stances within the same religious group. For example, in Myanmar, there were Buddhist monks who supported the military coup in 2021, but there were other monks who demonstrated against the coup (Artinger and Rowand Reference Artinger and Rowand2021). Furthermore, the focus was only on the institutional level in this article; future studies could address the behaviour of national or local religious leaders – whether from majority or minority denominations – towards autocratic practices that are related to regime survival.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/gov.2023.39.
Acknowledgements
I would like to express my profound appreciation for Professor Paul Djupe for being my mentor throughout the process of writing this article since its first draft, as part of the APSA MENA Mentoring Program. I am also grateful to Professors Mohamed Safy El-Din Kharboush and Mazen Hassan for supervising this work as part of my PhD dissertation. Also deserving of my gratitude are the ICPSR teaching staff, Professors Shawna Smith, Brian M. Pollins and Robert Cooper for their assistance and insightful instruction during the ICPSR summer programme in 2022 at the University of Michigan in Ann Arbor. I am also grateful for the three reviewers’ feedback, which has helped strengthen the article and clarify its main contributions. I would also like to extend my sincere thanks to Professors Mahmoud Hamad, Rawia Tawfik, Ebtisam Aly Hussein, Alexandra Blackman, Daniel L. Tavana, Nermin Allam and Wayde Z.C. Marsh for their support and insightful feedback.
