Globalization and religious resurgence: a comparative analysis

Abstract

Why does religion continue to emerge as a flashpoint in the age of secularization? Although models of religious resurgence suggest that religious cleavages are more prominent in the modern era, other models continue to show declining religious involvement. What is needed is a theory that can observe both resurgence and secularization at the same time. I argue that globalization—and the flow of people across borders, in particular—provokes a religious backlash at the societal level due to its secularizing effects. As the public is exposed to new and diverse religious traditions, religiosity declines; as a result, however, religious practitioners become more aggressive toward other religious groups. I test this theory using data on globalization, religious discrimination, and religious practice. I find that types of globalization dealing with the flow of people and information across borders have an outsize effect on societal religious discrimination, or SRD. This effect, however, is contingent on a decline in religious practice. This study suggests that religious resurgence can take place in secularizing environments, and that both resurgence and secularization share root causes.

Introduction

As states continue to secularize, why has religion emerged as one of the most prominent political flashpoints? Older theories of secularization suggest that religious practice should decline in relation to human development (Gaskins et al., 2013; Gill, 2021). Although long-standing trends in state religious policy (SRP) have drastically decreased the popularity of such theories, scholars continue to find support for a less-sweeping trend toward secularization in modernizing environments (Fox, 2016; Stolz, 2020; Dhima and Golder, 2021). But while religious practice has declined in certain contexts, religious minorities have faced increased discrimination by both government and societal actors (Fox, 2013, 2020; Fox et al., 2018). What is needed, then, is a theory that can observe both processes at the same time.

I argue that refocusing the question of secularization from modernization to globalization can provide key insights for the study of religious resurgence. Qualitative and theoretical research suggests that the backlash against globalization may lead to a form of “counter-secularization” in which religious movements encourage hostility toward other belief systems (Barber, 1995; Fordahl and Ragnarsdóttír, 2021). Although previous quantitative research has linked the religious resurgence to globalization, it has done so only at the government level and has not tackled the question of counter-secularization in society (Bloom et al., 2013). Where secularization and resurgence have been examined at the societal level, they have not been examined in relation to counter-secularization or socio-economic structures (Fox et al., 2021). Furthermore, previous studies have treated this relationship as effectively uniform across diverse geographic regions and political environments.

Based on literature surrounding counter-secularization, I argue that globalization facilitates secularization not through modernization mechanisms such as economic growth, but through increased contact between religious groups. As members of the dominant religion are exposed to diverse religious traditions, primarily through immigration and international communication, the dominant religion becomes less influential in society. As a result, however, religious actors may encourage hostility or discrimination against other religions in order to protect their status. If that is the case, then the relationship between religious resurgence and globalization should occur as a bottom-up, rather than top-down process. Furthermore, this relationship should be stronger in geographic regions where the social aspects of globalization have contributed to the process of secularization. In effect, I argue that that globalization does not drive religious resurgence in spite of secularization, but because of it.

I test this theory using data from the Religion and State (RAS) project, World Values Survey (WVS), and KOF Globalisation [sic] Index. I find that globalization exacerbates societal religious discrimination, or SRD, over time, but this effect is not uniform across types of globalization or geographic regions. Globalization types dealing with standards of living such as economic trade or finance show lower and less consistent effects, while those dealing with social issues, such as immigration and international communication, show the largest and most consistent effects. This is particularly true in regions that have been named hotbeds of religious resurgence. In addition, I find that various measures of religious practice decline in relation to the same mechanisms. Finally, the effect of globalization on SRD appears to be dependent on secularization, suggesting that secularization is a link in a chain—not something that can be viewed in isolation.

This study proceeds as follows: first, I review literature surrounding secularization theory and the religious resurgence. Here, I outline key findings that continue to show certain types of secularization despite a scholarly consensus that religion is resurgent in modern politics. Second, I discuss the literature surrounding counter-secularization and globalization, highlighting prior theoretical and qualitative contributions while further demonstrating key gaps in quantitative literature. Next, I present a research design and discuss the results. The first set of results compares the effects of different types of globalization on SRD, across varying regions. The second set of results compares the effect of each type on aggregate levels of religious practice. The third set of results demonstrates that the effect of key globalization types on SRD is contingent upon declining religious practice. Additional findings in the Appendix compare alternate hypotheses. Following a presentation of the results, I conclude with a summary of the findings and suggestions for further research.

Secularization and resurgence

A common narrative suggests that secularization theory—the theory that modernization reduces religious practice—held a hegemonic influence over the social sciences until the 21st century (cf. Swatos and Christiano, 1999; Gorski and Altinordu, 2008). Since then, however, a scholarly consensus has emerged regarding the unexpected resurgence of religiosity and the failure of secularization theory (e.g., Fox, 2016). In this telling, an increase not only in religious practice, but also in SRP, points to the continued importance of religious cleavages. If scholars conceive of secularization as an increasingly strong separation of religion and state, then states' increasing involvement religious institutions and practices points to renewed competition between religious and secular forces (cf. Fox, 2013, 2019a, 2019b, 2020).¹

Such analyses discuss many aspects of SRP, including both the state's support for religious institutions as well as the regulation of religious institutions. The most common focus, however, is religious discrimination—policies directed toward a state's religious minorities but not its majority religion. Because it represents a tangible cleavage between religious groups, the rise in discrimination directly challenges secularization theory (Fox, 2013, 2020). Existing studies using data from the RAS project, have explored both government religious discrimination, or GRD, and SRD (cf. Fox et al., 2018, 2021). While the former describes the prevalence of discriminatory policies in government, the latter describes the prevalence of discriminatory practices in society. These studies demonstrate that such discriminatory practices have increased in recent years, in some cases even in tandem with declining rates of religious practice.

Nevertheless, proponents of secularization such as Inglehart and Norris (2004) remain a focal point for scholarly debate. Gaskins et al. (2013) find that although modernization decreases religious practice, it also increases conservatism among those who continue to practice. Such findings indicate the possibility that religious actors and religious institutions may continue to exert influence even as the overall number of practitioners declines. Still, findings relating to modernization tend to produce mixed results depending on how secularization is defined (Stolz, 2020). Although increased standards of living at the country level do appear to reduce rates of religiosity, the same results do not hold for individuals (Höllinger and Muckenhuber, 2019). In other words, as with comparisons of secularization and resurgence, government, society, and individual all appear to react to these stimuli differently.

Crucially, additional findings by Dhima and Golder (2021) and Lüchau (2014) further demonstrate that religious self-identification has not declined in relation to modernization, although institutional involvement in religion has.² In short, the Gaskins, Golder, and Siegel model, or GGS, continues to find support for a particular type of secularization theory; despite this, it is often taken for granted that secularization theory has been discredited due to the upsurge of religious policy-making in the 21st century. The RAS model continuously finds support for an increase in religious policy-making even in secularizing environments.³ I have illustrated these two models in Figure 1.

Figure 1. RAS and GGS models.

Despite the apparent contradiction between the two, the GGS and RAS models have not been explored in relation to each other. The GGS model is largely concerned with microfoundations—behaviors by individuals—while the RAS model deals with macro-level structures—institutions. What the two literatures, secularization and resurgence, have failed to do is to observe the two phenomena at the same time—not as polar opposites but as two sides of the same coin.

In this study, I seek to further explore another area of literature that can potentially bridge this gap. Rather than modernization producing secularism, I suggest refocusing the framework toward globalization. The shift from modernization to globalization as a causal mechanism enables an examination of both secularization and resurgence in relation to different aspects of social, economic, and political change. As I will discuss, this framework has gone under-explored in quantitative analysis, and the research that has been done has either concentrated on theoretical or qualitative analyses or has yet to receive significant extensions from its foundational literatures. In the following section, I review this body of literature, using the terminology “counter-secularization” to describe the broader framework.

Counter-secularization and globalization

Globalization brings disparate groups into contact with each other. When new interactions between groups emerge, they can take on both hostile or amicable characteristics depending on the context of these new interactions. The long-standing “contact hypothesis” suggests that if groups interact on a level playing field, they will become more accepting of each other. By the same token, the “group threat” hypothesis suggests that if members of one or both of the groups fear for their status, they will not be able to find common ground.⁴ The effect of globalization, then, largely depends upon the conditions of contact between groups (Kaya and Karakoc, 2012). In a religious context, however, there is reason to believe that it has had both effects. I argue that globalization facilitates secularization according to two principles: first, it allows religious practitioners to experience life outside of their own community; second—as a byproduct of the first principle—it exposes religious practitioners to different faiths.

The “contact hypothesis” has been discussed at length in relation to religion and religious freedom. Both Berger and Zijderveld (2009) and Grim and Finke (2010) argue that globalization naturally leads to religious diversity. These authors suggest that such diversity will reduce religious tensions. If societal actors fear for their status, however, this line of reasoning may not hold. In a study of non-state actors, Feinberg (2020) demonstrates that discrimination against religious minorities is more likely when there is opportunity, distinguishability, stimuli, and organization.

Globalization can provide each of these, especially opportunity and stimuli. If newfound religious diversity is also associated with a decline in the status of the dominant religious group, the dominant religious group may seek to re-assert itself. According to Gill (2008), religious diversity lends itself to greater religious freedom, which undermines the dominant religious group's influence. Both Gill and Grim and Finke further argue that the lasting influence of dominant religious groups under threat may cause a “tyranny of the majority” effect in which the dominant religious group pressures government into adopting restrictive policies against their competitors. Such grievances are borne out when examining the dynamics of religious communities and religious actors' critiques of globalization, secularism, and religious diversity.

Religious communities revolve around kinship ties and dense social networks, hence the term “parochial” (cf. Fukuyama, 2001). Globalization, by contrast, encourages a vast network of cosmopolitan social organizations that transcends local identities—and often treats them as irrelevant (cf. Dryzek, 2012). The freedom of movement and the free flow of ideas across borders allow individuals to remove themselves from their traditional communities, which—at least in theory—revolve more heavily around religion (Hochschild, 2003, 2006). Critics contend that removing individuals from a communal context homogenizes cultures (Deneen, 2018). This further extends to critiques that globalization pressures non-Western cultures to integrate into Western systems (Mazrui, 1998).

Among religious ideologues, critiques of globalization and secularism alike outweigh critiques against modernization. Evangelical preacher Robertson (1991), for example, decried globalized institutions as a means of separating individuals from their parochial communities and exposing them to alternative lifestyles. Similar complaints have been made by the Ayatollah Khomeini (1970), Jewish extremist Kahane (1987), Bin Laden (2016), and Myanmar's infamous extremist monk Ashin Wirtahu (cf. Foxeus, 2019). Although these thinkers possess a wide range of ideological and religious identities, they each share common critiques of globalization—as well as invocations against religious minorities.

In each of these examples, ideologues accused religious minorities of using globalized institutions to water down the majority population's religious identity while preserving their own unique identities. Because religious identities traditionally have exclusive characteristics, exposure to other faiths can open a dominant religion to new lines of questioning.⁵ Robertson, for example, argued that “globalists” sought to “erase the minds of the students, to tell them that everything is relative,” adding that “the million gods of the Hindus, and the Allah of the Muslims is not the God of Jacob” (217–231). Khomeini complained that the secular “agents of imperialism… are not converting [Muslims] into Jews and Christians; they are corrupting them, making them irreligious and indifferent” (79). As religious ideologues draw parallels between religious minorities, secularism, and globalization, they also broadly conflate religious out-groups. Rather than targeting specific religious minorities, these actors tend to view all forms of religious diversity as a threat to the status of the dominant religion.

These theoretical foundations lay the groundwork for two simultaneous phenomena. Secularization may follow from the contact hypothesis; that the influence of religious choice and the removal of barriers between religious communities allows much of the public to “opt-out” of religion. However, religious actors and practitioners who see their communities shrinking may instead focus their ire on the influx of other religions due to the perception of threat. The two processes—secularization and resurgence—can occur simultaneously because religious discrimination can occur within the context of a secularizing environment, as part of a backlash against secularization itself.

Globalization matters for both secularization and resurgence because it is the driving force behind the two impulses of “contact” and “group-threat.” It is functionally distinct from “modernization” as a causal mechanism because the two concepts measure different phenomena. In its broad strokes, globalization refers to the institutional links across national borders (Potrafke, 2014). Although these can include economic indicators related to standards of living, they focus on a country's openness to the flow of goods, services, people, and information. Because these phenomena are regulated by a country's laws, this method of examining social, economic, and political change speaks to institutional environments more than indicators of modernization. They therefore consider broad mechanisms of change in terms of standards of living and the relationship between individuals, society, and government.

Although a substantial body of literature discusses the role of globalization in facilitating religious cleavages, these discussions have historically focused on theoretical foundations or “big-picture” clashes in the international arena. According to this camp, evidence suggests that the religious sects that are growing are the ones that reject progressivism (Berger, 1999). Sects or practices that reject progressivism act as forces of a “counter-secularization” process that often targets religious minorities as an out-group (Karpov, 2010).⁶ Barber (1995) argues that these ideologies arise as a reaction to globalized institutions, in an effort to protect decentralized communities.

Other scholars famously see modern religious tensions as a product of civilizational fault lines (Huntington, 1996). These scholars view the “clash of civilizations” as a form of “cultural resurgence” as globalization brings religious and national groups into contact with each other (Collet and Inoguchi, 2012). Brubaker (2017), for example, argues that widespread immigration from non-Christian countries has led many secular politicians in the “West” to embrace a secularized Christian identity. Although Western Europe is a largely secular culture, Christian religious identity serves as the dividing line between “Western” and “Islamic” civilization. Both Huntington (1996, 201–202) and Kinnvall (2004) argue that a similar process has taken place in the Muslim-majority world, as globalization has facilitated greater contact between Christianity and Islam; as proselytizing religions, each civilizational identity experiences newfound tension due to this increased contact.⁷

Connecting the disparate strands of this literature, Fordahl and Ragnarsdóttir (2021) interpret the development of aggressive religious identities as part of a “post-secularization” process; here, globalization drives newfound religious practices in secularizing environments—most critically, forms of religion that seek to exclude members of other religions from the body politic. This work, however, is tailored to a particular case study and has yet to be utilized in a broader, comparative context.

Discrimination is the logical outcome of these exacerbated cleavages, because it is a frequent means of reasserting collective identity given threats to group cohesion (Brewer, 1999). Some prior work has incorporated elements of discrimination, globalization, counter-secularization, but thus far these studies have not received proper follow-ups. Critically, Bloom et al. (2013) argue that religious resurgence is a response to “globalization threat”—the sense of threat that the public feels due to globalizing institutions' potentially homogenizing effect on dominant cultures. In this model, the public's fear of globalization manifests itself as opposition to religious pluralism, and the public's opposition to religious pluralism leads to SRP as a response.

Bloom, Arikan, and Sommer's model tests this theory at the government and individual levels, and not at the societal level. Given that previous studies such as Fox (2020) have named SRD as a precursor to GRD, this is an important step to be filled.⁸ The societal level, acting as an intermediary level between government and individuals, is key to bridging the GGS and RAS models. However, these findings do not differentiate between distinct geographic regions, political environments, or religious settings, and their definition of globalization relies on older data that is less specific than modern data. Subsequent studies have largely focused on the findings' implications for religious identity or even the quality of democracy as a whole (cf. Fox et al., 2019; Jerabek, 2022). More recently, Zhiang (2020) provides a direct follow-up to Bloom, Arikan, and Sommer's findings, but focuses on religious policies and does not examine the societal level.

In this study, I offer a ground-up model of counter-secularization. In doing so, I seek to reconcile GGS models with RAS models. I argue that specific mechanisms of globalization—particularly migration and information—both reduce institutional involvement in religion as well as spark backlashes from religious practitioners. In other words, while religious practice is indeed declining, the remaining religious practitioners are growing more aggressive toward other religions.

If this is the case, then religious discrimination should rise in conjunction with specific types of globalization and not others, while religious practice should fall in relation to the same types. We can expect these globalization types to be those related to increased social contact across borders, particularly immigration and information. The hypotheses are as follows:

Hypothesis 1 (H1). Social types of globalization will increase SRD over time.

Hypothesis 2 (H2). Social types of globalization will decrease religious practice over time.

At the same time, the theory above suggests that globalization exacerbates SRD because of secularization—that is, that these phenomena are intricately linked. To that end, a third hypothesis examines the effect of globalization in conjunction with declining religious practice:

Hypothesis 3A (H3A). The effect of social globalization types on SRD requires a concurrent decline in religiosity.

This hypothesis does not suggest that societies with high levels of religiosity are more tolerant of religious minorities; rather it suggests that globalization produces a change in the status of religion—and religious actors—and therefore invites a backlash. Therefore, a parallel hypothesis must be considered:

Hypothesis 3B (H3b). The effect of religiosity on SRD will be dependent on the level of social globalization.

If declining religiosity is a necessary condition for the relationship between globalization and SRD, this could imply that higher levels of religiosity reduce SRD. I suggest, however, that this is only true within the context of globalizing societies. Examining the effect of these variables in relation to each other will demonstrate that their effects are contingent upon each other. Rather than independently affecting SRD, these variables can be viewed as part of a series of events.

Data and methods

I use a series of ordinary least-squares (OLS) model to gauge the impact of globalization on SRD and religiosity by country-year. In order to observe within-unit change over time, I have deployed country-year fixed effects (cf. Fox et al., 2017). Both the primary dependent variable as well as two control variables are taken from the RAS project round 3, or RAS3, which encompasses 183 countries spanning the years 1990–2014. RAS data are also coded by region. Although initial findings by Bloom et al. (2013) include all countries, more recent studies, such as Fox (2020), use these regions to group results, allowing for a more reliable analysis. Thus, regarding H1, I conduct tests according to regional groupings rather than pooled into one sample. These are further presented according to varying categories of the independent variable. The first hypothesis is examined on a comparative basis. As I will discuss, this is not feasible concerning H2 or H3 given the limited availability of data. For H3—both A and B—I further include an interactive effect between causal mechanisms—globalization types and rates of religious practice—to establish a causal pathway rather than simultaneous causation. I limit these tests to types of globalization that have established effects according to H1 and H2 (Table 1).

Table 1. Descriptive statistics for dependent variables

The first dependent variable is Societal Religious Discrimination, or SRD. This variable, which appears in the RAS3, measures discriminatory practices against minority religions at the societal level. Although other RAS3 variables typically describe regulation of religious practices, SRD is one of the few based on actions, not policies. As such, it tends to change more from year-to-year than other variables in the dataset.⁹ It also describes acts of physical violence that cannot be captured by government policy alone.

The RAS3 measure of SRD is an index composed of 27 items encompassing a wide range of discriminatory behaviors by non-state actors toward religious minorities.¹⁰ Each item is rated on a scale of 0–3, 0 indicating that the discriminatory behavior is not present, and 3 indicating that it is frequent toward all religious minority groups. Discriminatory actions include a wide variety of items, including types of economic discrimination, hate speech, non-violent acts, and violent acts. Some actions included in the index involve attacks or vandalism of property, while others involve people with visible affiliations to the religion.

The overall index is the sum of each item.¹¹ As a result, the potential values for the SRD index range from 0 to 81. A value of 81 would indicate that all forms of religious discrimination are widespread and affect all religious minorities. In actuality, however, no country-year features a score higher than 62—a distinction belonging to Egypt from 2009–2014. A score of 62 suggests that most forms of SRD are widespread and affect all religious minority groups rather than particular groups, but not all forms of SRD are practiced (Akbaba and Fox, 2019). Small SRD scores, in the range of 3–7, typically imply that religious groups have been selectively targeted on a limited basis. Lower scores of SRD are more common outside of Europe, North America, and the Middle East–North Africa, although in some countries such as the United Arab Emirates, SRD is quite low even though levels of government policy-making on religion are quite high. Latin America and sub-Saharan Africa show the lowest levels of SRD.

The other dependent variables concerning indicators of religiosity are from the WVS (Inglehart et al. 2014). The first is the aggregate percent of WVS respondents attending religious services at least once a month (cf. Cingranelli and Kalmick, 2019; Fox et al., 2021). It covers the first seven waves of the project, encompassing all available data as of November 2021. In order to observe within-unit change over time, I used linear interpolation to fill values between years. Observed values range from just over 1% to 98.383%, with the average at 59.587. Although service attendance is frequently used to gauge religiosity, critics contend that this primarily describes Western religious practices rather than religious practice as a whole (Verghese, 2020). To account for the potential disparity, I have also included the rate of regular prayer. This is the percent of WVS respondents who report praying at least several times a week—the next lowest rating being “only at religious services.” This variable ranges from 0.8% to 95.62%, with an average of 41.134 (Table 2).

Table 2. Globalization types

The independent variables consist of several measures of globalization, divided into different types. These measures are drawn from the index categories in the KOF Globalisation [sic] Index (Dreher, 2006; Gygli et al., 2019). Each index is based on a series of variables related to an aspect of globalization. Taken together, each index can be compiled into a single value indicating the level of globalization in a given country. At the same time, the data also group these variables into specific types: economic, social, and political. The economic index of globalization can be further broken down into financial and trade-related variables, while the social index can be broken down into interpersonal, informational, and cultural variables.

For the purpose of this study, interpersonal and informational types of globalization are the most important. The former describes the flow of people across borders. It is compiled from a number of measures that broadly describe freedom of movement, such as the number of foreign-born residents, strength of international tourism, and transfers of goods and services, along with a handful of other items. The latter describes the public's access to information, relating both to internet and television as well as items such as freedom of press and relative frequency of international patents. These concepts are particularly important when analyzing the resurgence of religious identity, because they directly challenge dominant religions in two ways: first, by exposing the public to a diverse array of outsiders with a diverse set of practices, and second, by introducing the general public to narratives that run counter to dominant ideologies, including religious dogmas.

The broader index compiles a total of 40 variables into a handful of categories describing types of globalization.¹² I have detailed the composition of these types in Table 2. Each index is scaled from 0 to 100. For H1, the effect of each type of globalization is presented according to regional groupings.¹³ In other words, I have presented nine different types of globalization across six different regions to examine the relationship between globalization and religious resurgence on a fine-grained level. This is not possible for H2 and H3, due to the relatively small number of country-year observations in each region.

In addition to the variables above, I have deployed a standard set of control variables previously utilized by Fox (2013, 2015, 2020) and Fox et al. (2021), excluding time-invariant variables rendered superfluous by fixed-effects models. For controls, I have included religious fractionalization, GDP per capita, population size, democracy scores, violent minority actions, and GRD. Religious fractionalization is a Herfindahl–Hirschman index of the Religious Characteristics of States dataset, based on the provided taxonomy (Brown and James, 2017). I use this measure in place of minority percent (cf. Fox, 2020; Fox et al., 2021) because I am using countries as the unit of analysis rather than minority groups. GDP per capita and population numbers appear in the World Bank's (2021) World Development Indicators, logged for skew. Democracy scores are taken from Polity V; they range from −10 to 10, with higher scores indicating greater democracy (Marshall and Gurr, 2018). Finally, violent minority actions and GRD are both drawn from the RAS3. Violent minority actions include violence by religious minorities against majority religions; I have excluded values for minority–minority discrimination as it is not reliable at the country level (Fox et al., 2018). GRD is calculated in the same vein as SRD, using a total of 36 categories.

Given that societal religious cleavages have previously been theorized as a source of restrictive government policies (Gill, 2008; Grim and Finke, 2010), most studies such as Fox (2020) control for SRD in tests of GRD but not vice-versa. Here, I control for GRD to account for cases where widespread religious observance and homogeneity are maintained through government coercion. This is particularly important for the Middle East–North Africa, where many states are more religiously homogenous than average, and GRD is significantly higher than SRD. To assuage any concerns regarding this decision, I have added an additional Appendix (Appendix 5) that compares results side-by-side.¹⁴

Results

Figure 2 compares the outputs of several OLS models using coefficient plots. Controls are suppressed for space. Full regression tables can be found in Appendix 1.

Figure 2. Coefficient plots for H1.

In the first set of tests, I present side-by-side results demonstrating the impact of several distinct types of globalization on SRD, grouped by each region codified in the RAS. These tests include each of the nine coded types of globalization—aggregate and disaggregate alike—across six regions. The combined set of 54 tests detailed here demonstrates significant variation across types of globalization and regions of particular impact, allowing for greater precision in exploring the relationship on a comparative level.¹⁵

Although largely consistent with H1, results vary considerably by region. Western democracies and the Middle East–North Africa each show the highest effect sizes, particularly as it concerns interpersonal globalization. In Western democracies, the range of interpersonal globalization predicts a 15-point increase in SRD; in the Middle East–North Africa, it predicts a nearly 20-point increase in SRD. In Western democracies, this relationship represents an increase from about 6 points on the SRD scale to about 14—representing a rise from 45th percentile to 77th within the region. The same numbers are even starker for the Middle East–North Africa. Here, the total range of interpersonal globalization values predicts a rise in SRD from roughly 3 to 22—or 25th percentile to 90th. This means that over the range of interpersonal globalization, countries in these regions move from relatively low- or medium-range levels of SRD to quite high levels. Both effects are significant at the 0.01 level (p < 0.01).¹⁶ These effects are illustrated in Figure 3.

Figure 3. Illustrative effects of H1.

Of all the types of globalization, the “interpersonal” type shows the most consistent pattern across regions, demonstrating a significant effect in every region except for Latin America (p < 0.01). Here, however, R ² values prove substantially lower, with the highest additional value reaching only 0.267 in sub-Saharan Africa. Thus, outside of Western democracies and the Middle East–North Africa region, the effect of interpersonal globalization only accounts for about one-fifth to one-fourth of the variation in SRD. In terms of the variation in globalization types themselves, informational globalization also demonstrates a robust impact, appearing significant at the 0.01 level (p < 0.01) in every region except for Latin America and sub-Saharan Africa. Here, too, the effect sizes are larger in Western democracies and the Middle East–North African regions, predicting rises in 8 and 6 points, respectively.

Other results are less consistent: trade globalization is only significant in Eastern Europe; political and cultural globalization are only significant in Asia. The effect size of interpersonal globalization, while still significant, is less than half of that in Eastern Europe and Asia as it is in Western democracies and the Middle East–North Africa. In Latin America, meanwhile, all globalization types are shown to exert a negative effect on SRD—significant at the 0.01 level (p < 0.01)—with the exception of interpersonal and cultural types. This sole holdout to the overall trend likely speaks to the role religious institutions played in Latin America's wave of democratization (cf. Huntington, 1991). It also indicates that this relationship is more heavily indebted to the structural interaction between religion, state, and society, rather than the dogmatic aspects of specific religions.

Ultimately, these tests indicate contextual support for H1; the relationship between globalization and religious cleavages as observed through SRD is contingent on a number of factors. First, specific types of globalization demonstrate outsize impact. Second, this impact varies considerably by region. Third, the importance of this impact itself also varies by region. Beyond particular cases of interest, other structural elements remain at work that are not captured by these models. Further research is thus needed to parse the confounding factors in regions with lesser impact, or particular structural factors that render particular regions especially vulnerable to this phenomenon.

These tests cannot and do not attempt to answer the question of whether it is religious or secular actors perpetuating SRD. Indeed, the literature surrounding counter-secularization suggests that this distinction is not absolute. Nevertheless, I have sought to demonstrate a potential path forward in this area by comparing the relationship between SRD and globalization types to religiosity as whole. To this end, Figure 4 contains coefficient plots for H2. Full regression tables are again found in Appendix 1. In this case, I have not divided results by region due to an inefficient number of groupings in each. Full regression tables are available in Appendix 1.

Figure 4. Coefficient plots for H2.

Results indicate that aggregate globalization decreases both service attendance and prayer rates; but that specifically social, and interpersonal types in particular, demonstrate the largest deleterious impacts. Over the range of interpersonal globalization, service attendance rates drop nearly 18%, corresponding to about 66th percentile to 43rd; prayer rates, meanwhile, fall nearly 43%—from 72nd percentile to just 29th. Both effects are significant at the 0.01 level (p < 0.01).¹⁷ Social globalization, and interpersonal globalization in particular, exerts a consistent effect across SRD, service attendance, and prayer. In other words, expressions of religiosity often respond to the same mechanisms as SRD. Every type of globalization that decreases service attendance also increases SRD, and every type of globalization except cultural decreases regular prayer.

But are these effects occurring simultaneously, or are they conditional on one another? Fox et al. (2021), for example, suggest that the inverse relationship between secularization and discrimination may be due to secular actors' hostility to religious actors. Regarding H3 (A and B), I have prepared an additional test that seeks to gauge this question. Although the limited availability of data precludes a fine-grained comparison like that of H1, these tests include an interaction term between globalization and religiosity. This means that the effects of globalization and religiosity on SRD are each being viewed in relation to each other.

These tests are illustrated in Figure 5, depicting H3A and H3B respectively. Full regression tables are found in Appendix 1. According to the results for H3A, depicted on the left-hand side, the effect of interpersonal globalization on SRD is contingent on a decline in religiosity—in other words, secularization. This is to say that when interpersonal globalization is not accompanied by a decline in religiosity, it does not predict SRD. These results are significant at the 0.01 level (p < 0.01). At high levels of religiosity, interpersonal globalization is actually weakly associated with a decrease in SRD (p < 0.05). In short, interpersonal globalization—the flow of people across borders—does not exacerbate religious cleavages unless it is also accompanied by a loss of religiosity in society writ large.

Figure 5. Test for H3.

The same logic holds when analyzing H3B, the relationship between religiosity and SRD, as depicted on the right-hand side. Findings thus far suggest that secularization exacerbates SRD; this means that the relationship between discrimination and religiosity should be inverse. This is only true, however, at high levels of interpersonal globalization. When interpersonal globalization is low, religiosity is in fact associated with an increase in SRD. In other words, while the relationship between globalization and SRD is contingent on declining levels of religiosity, the inverse relationship between religiosity and SRD is contingent on high levels of globalization. This suggests that neither declining religiosity nor globalization exacerbate SRD on their own. Rather, they form a sequence of causal events.¹⁸

In short, these findings offer some nuance to the hypotheses. Although these models suggest that globalization does increase SRD, they further suggest that this relationship is highly contextual. This is also the case with regard to the relationship between secularization and SRD. Globalization only exacerbates SRD insofar as it is accompanied by a decline in religiosity, while religiosity is only associated with a decrease in SRD if it is accompanied by globalization. At the same time, the variation in regional and typographical patterns suggests that this relationship should be further examined for structural elements that further condition results in one direction or another. Further study is needed to examine these structural elements on a fine-grained basis. Further study is also needed to disaggregate globalization types to their barest components to demonstrate highly specific relationships.

It is also worth noting that throughout these tests, the inclusion or exclusion of GRD rarely alters results. The fact that most model results are unaffected offers further support for previous theories such as Grim and Finke (2010) that suggest societal restrictions influence government restrictions, rather than the opposite. There are, however, caveats to this aspect, as discussed in Appendix 5.

Discussion

Although alternatives to secularization have grown increasingly prominent and popular, models such as the GGS continue to find support for a limited type of secularization. At the individual level, human development appears to reduce religiosity. Nevertheless, models of institutional structures like the RAS project have consistently shown a resurgence of religious activity at the country level. In recent years, academic discourse has increasingly moved beyond the push-and-pull between theories of secularization and resurgence. Nonetheless, scholars continue to find evidence of both theories. As such, these ostensibly contradictory results may in fact be two halves of the same whole. Reconciling these models is key to building a cohesive picture of religion, state, and individual in a globalized age.

Referring to a third body of theoretical literature—here, dubbed “counter-secularization”—this article sought to bridge these models, as a first step toward reconciliation. According to the literature of counter-secularization, the forces of globalization, rather than the process of modernization, facilitates religious cleavages by bringing disparate religious groups into contact with each other. As dominant religious groups perceive a threat from religious diversity, they seek to protect their status.

Although previous studies examined the relationship between globalization threat and SRP, this article examines the relationship between globalization and societal expressions of both religiosity and religious cleavages. Results indicate that globalization is indeed associated with both SRD, as an expression of religious cleavages, as well as a decline in aggregate levels of religiosity. Nevertheless, these findings also presented several conditions for this relationship: first, that this relationship is stronger in some geographic regions than others; second, that this relationship largely stems from interpersonal and informational globalization, rather than globalization of markets or of political institutions; and finally, that globalization only exacerbates SRD if religiosity is declining. In other words, the connection between globalization and resurgence is not absolute, and neither is the connection between secularization and discrimination; instead, they are links in chain.

It is worth noting that the elements of globalization that show the greatest effect are those that increase diversity and pluralism. Interpersonal globalization, which demonstrated the strongest and most consistent effects in a wide variety of contexts, encompasses the flow of people across borders, including broad points of contact ranging from immigration to tourism. The second strongest in this regard, informational globalization, speaks to a population's access to a global marketplace of ideas.

Further study is needed to explore these findings. A broader structural analysis entailing both individual-level as well as aggregate data may provide further insight as to how macrolevel trends impact microlevel phenomena. For a full view of this subject matter, an analysis must account for intermediary variables within a multistage causal chain. This article thus lays the foundation for a broader effort to integrate schools of thought concerning the relationship between religion and state in the modern world.

Indeed, there are many connections between these concepts that are worth exploring further. It is impossible, for example, to remove interpersonal and informational globalization from the context of the transnational marketplace (cf. Helms, 2024). It is also impossible to remove the transnational marketplace from the context of transnational political institutions (cf. Gygli et al., 2019). While social globalization appears to have an outsize effect, it must be recognized that social globalization can itself be part of a process that involves both political and economic elements. Beyond globalization itself, it is impossible to remove these developments from the context of religious movements which both aided and opposed such developments in both developed and developing economies (cf. Dreher, 2020).

These findings, then, suggest that it is also impossible to fully extricate the religious resurgence from a broader environment of secularization. The relationship between the two may in fact be highly nested in nature. Individuals, religious families, and even administrative units may yield distinct trends and phenomena relative to religious institutions, countries, geographic regions, and larger religious blocs. These effects may be relatively inconsistent in some stages of development, and relatively consistent in others. Relationships between these variables, then, should be viewed in context rather than in isolation. Although this study seeks to clarify and reconcile existing arguments, it is not an end point. Rather, it points in the direction of potential future lines of inquiry.