The Political Sources of Religious Identification: Evidence from the Burkina Faso–Côte d’Ivoire Border
The Political Sources of Religious Identification: Evidence from the Burkina Faso–Côte d’Ivoire Border
Introduction
Core Research Question: Under what conditions does religion become a salient social identity?
Global Context: An estimated 85\% of the world’s population claims membership in a religious group, yet the importance attached to religious identity varies considerably.
Variations in Religious Salience: For some, religion is a defining characteristic guiding politics and worldviews; for others, it's merely one social attachment among many, with little impact on attitudes and behavior.
Motivation for Study: Increased importance for social scientists to evaluate sources of religious motivation, given global events spearheaded by actors claiming religious piety (e.g., Afghanistan, Iraq, Nigeria, Central African Republic, Syria, Israel, Yemen).
Article's Thesis: Political events, particularly political elites' mobilization of religion for strategic purposes, can alter the importance people attach to their religious identities.
Contribution to Literature: While many instrumentalist studies focus on elite mobilization of social identities, fewer document the impact on how regular citizens self-identify, and fewer still explicitly focus on religious attachments.
Contrast with Traditional Accounts: This study emphasizes political sources, in contrast to traditional accounts that focus on:
Social and economic factors like economic development and secularization.
The supply of different religions and the competition between them.
The content of specific religious practices and belief systems.
Research Design and Methodology
Ideal Research Design: An experiment randomly assigning people to distinct political environments to compare religious identification (not feasible outside a laboratory).
Quasi-experimental Approximation: Exploiting the exogenous changes in political environments provided by the 19th-century partition of Africa.
Rationale: Colonial officials drew boundaries with little knowledge of local populations, leading to 'as-if random' assignment of populations to different political contexts.
Causal Inference: Differences in religious salience can be attributed to country environments, assuming initial characteristics were similar and controlling for other differing factors.
Study Area: The Burkina Faso–Côte d’Ivoire border in West Africa.
Benefit: Focusing on populations living very close to the border allows holding constant many social, economic, cultural, and historical factors while varying the causal variable of interest (the national political environment).
Arbitrary Border Confirmation:
Origin: Established in the 1880s as administrative district (cercles) boundaries in French Sudan.
Path: Runs west to east from the Bani and Léraba Rivers to the Black Volta River.
Demarcation Method: Primarily followed natural geographic features (seasonal rivers, streams) or simply connected dots. Maurice Delafosse described it as drawn 'in a sensible fashion […] to the East and Southeast'.
Pre-existing Similarities:
Huillery's data (2006) on socio-economic status and colonial investment in the early 20th century suggest cercles on both sides were very similar.
Delafosse's diaries (1912) corroborate this, noting no dissimilarities in local communities' habits of dress, scarification, circumcision, or weaponry.
Data Collection:
Survey: Administered to approximately 200 respondents across two pairs of research sites just astride the border.
Sites: Rural villages (Boussoukoula, Burkina Faso; Kalamou, Côte d’Ivoire – approx. 7 kilometers apart) and urban market towns (Niangoloko, Burkina Faso; Ouangolodougou, Côte d’Ivoire – approx. 50 kilometers apart).
Sampling: 50 respondents recruited in each site via random sampling, stratified by age and gender.
Languages: Dioula and/or French in urban sites; Lobiri (and occasionally French) in rural locations.
Response Rate: Achieved a high 93\% response rate, mitigating concerns of systematic bias despite a relatively small sample size.
Power Calculations: A total sample size of 150 (seventy-five per side) is sufficient to detect an effect of the reported magnitude (alpha = 0.05, standard power = 0.8).
Characteristics of the Sample
General Similarities: Respondents from Burkina Faso and Côte d’Ivoire were, on average, indistinguishable across most measurable characteristics (Appendix Table A1).
Key Differences Identified:
Religious Composition: Burkinabé respondents were significantly less likely to be Muslim (37\% versus 66\%) and more likely to be Catholic (36\% versus 11\%) than Ivorian counterparts.
Migration Rates: Respondents in Burkina Faso had lived in their current location ~25\% longer (avg. 25.2 vs. 20 years). They were twice as likely to be born in their survey village/town (52\% vs. 26\%) and about three times less likely to have migrated from across the border (13\% vs. 35\%). This aligns with Côte d'Ivoire's historical labor migration policies and Burkina Faso being a migrant-sending country.
Issue of Endogenous Sorting: Different migration rates raise the possibility that people moved for reasons related to religious attachment, potentially biasing results.
Measuring and Explaining Variation in Religious Salience
Survey Questions on Religious Salience:
'For you, what identity is most important?'
'After that, what identity would you place in second position?'
'Could you marry a person of a different religion?'
'To whom do you feel closer: a person of your country who is not of your religion, or a person of your religion who is not of your country?'
Striking Cross-Border Differences (Table 1, using conservative cluster mean analysis):
Religion as Most Important Identity: Ivorian respondents were almost three times more likely (27.8\%) than Burkinabé (10.0\%) to rank religion as their primary identity (difference of 17.8\%,
p < 0.10).Nationality as Most Important Identity: Ivorian respondents were far less likely to mention nationality (25\% versus 42\% for Burkinabé).
Religion Among Top Two Identities: Ivorian respondents ranked religion among their top two identities at a much higher rate (49.7\%) than Burkinabé individuals (23\%) (difference of 26.7\%,
p ext{ between } 0.10 ext{ and } 0.15).Willingness to Marry Across Religious Lines: Ivorian respondents were less likely (57.1\%) than Burkinabé counterparts (75.0\%) to say they could marry a person of a different religion (difference of -17.9\%,
p ext{ between } 0.10 ext{ and } 0.15).Proximity to Co-religionists vs. Co-nationals: Ivorians were less likely (40.3\%) to favor co-nationals over co-religionists compared to Burkinabé respondents (77.3\%$) (difference of -37.0\%,
p < 0.10).
Addressing Endogenous Sorting (Migration):
Problem: Migrants were more likely to prioritize religious identities (28.8\% of migrants vs. 15.7\% of non-migrants).
Conservative Approach 1: Country of Origin: Using birth country as the independent variable (like intention-to-treat).
Result (Table 2, Row 2): The cross-country difference was reduced but still significant: 15.9\% of Burkina Faso-born respondents prioritized religion vs. 23.2\% of Côte d’Ivoire-born respondents (difference of 7.3\%,
p < 0.10).
Approach 2: Country of Residence as of Political Treatment Start (2000/1993):
Result (Table 2, Rows 3 & 4): Differences persist. Residence in Côte d’Ivoire as of 2000 or 1993 still showed a significantly higher likelihood of listing religion first (e.g., 11.2\% difference for 2000 residence, p < 0.10).
Conclusion on Sorting: Cross-border differences in religious salience are not likely due to individuals sorting themselves across the border based on religion.
Explaining the Differential Salience: Ruling Out Alternative Explanations
Compound Treatment Problem: Acknowledges difficulty in isolating causal factors due to co-variation of religious demographics, political institutions, economic development, etc., at the country level.
1. Religious Demographics:
Hypothesis: Since Côte d'Ivoire has a higher proportion of Muslims, and Muslims tend to prioritize religion more (25.7\% vs. 10.6\% for Catholics), this could drive the observed differences.
Rebuttal: When the sample is limited to Muslim respondents, the cross-border difference in religious salience remains statistically significant (p=0.088 by cluster mean; p=0.01 at individual level), indicating the country effect dominates group membership.
Protestants: Also showed high prioritization of religion (25\%), but their small and balanced share across countries made their role less significant.
2. Wealth and Secularization:
Secularization Theory: Increased wealth and stability lead to a decline in religious salience, while 'existential insecurities' (poverty, hunger, disaster) maintain it.
Rebuttal: This study found the opposite of the predicted pattern. Religious identities are stronger in Côte d’Ivoire, despite its historically higher per capita incomes and slightly higher living standards among Ivorian respondents in the study area (Appendix Table A1). Demographic and Health Surveys also suggest similar socio-economic terms in border regions.
3. Quantity/Quality of Religious Institutions and Competition:
Hypothesis: Greater opportunities to practice faith, or increased inter/intra-religious competition from more institutions, could increase salience.
Rebuttal (Quantity): Interviews with religious leaders revealed nearly identical supplies of religious institutions in each site pair (Table 3).
Rural villages (Boussoukoula/Kalamou) each had two mosques, one Catholic Church, and one Protestant center.
Urban towns (Niangoloko/Ouangolodougou) had comparable total Muslim worship space (different structures) and identical numbers of Christian churches.
Rebuttal (Quality/Activity): Comparisons of build dates (a proxy for quality/longevity) showed strong similarities across site pairs.
First urban mosques appeared in the 1880s. Protestant churches in the 1930s, and large Catholic churches in the early 1940s.
Rural mosques appeared in 1984 and 1990; Christian churches only within the decade prior to research.
Attendance rates were also comparable relative to community sizes.
Rebuttal (National-level Pluralism): While Côte d’Ivoire's Religious Fractionalization score (0.755) is slightly higher than Burkina Faso's (0.580), both are in the top third worldwide, suggesting a difference of degree, not kind. Furthermore, national-level pluralism should manifest in local practices and worship places, which was not observed.
The Political Mobilization of Religion in Côte d’Ivoire (Preferred Explanation)
Historical Context:
Houphouët-Boigny Era: Religious and ethnic differences were largely mitigated through cooptation.
Post-1993: After President Houphouët-Boigny's death, politicians strategically exploited ethnic and religious differences for power, notably through the xenophobic policy of Ivoirité (Ivorianness).
Politicization of "Ivoirité":
Ivoirité rhetoric, while ostensibly ethno-national, translated at the street level to 'foreigner' meaning 'Muslim'.
This created deep fissures between Southern (largely Christian) and Northern (largely Muslim) Ivorians, as many Northerners were foreign migrants from predominantly Muslim countries.
Escalation to Conflict:
A coup in late 1999, disputed elections in 2000, and a 2002 rebellion divided the country.
Hostilities continued even after the official peace agreement in 2007.
The 2000 elections and subsequent violence are generally considered the start of the conflict, though identity divisions began after Houphouët-Boigny's death in 1993.
Elite Mobilization Examples:
Then-current President Alassane Ouattara (a northern Muslim) claimed he was prevented from running in 2000 'because I’m a Muslim'.
Former President Laurent Gbagbo (a southern Christian) labeled opposition forces from the North as 'the Ivorian Taliban'.
The conflict involved the burning of churches and mosques and the murder of religious leaders.
Both Gbagbo and Ouattara explicitly used religion to gain political legitimacy and support during the crisis (e.g., Gbagbo lobbied US evangelical Christians; Ouattara rallied the Superior Council of Imams).
Mechanism: These strategies align with the exploitation of the institutional and moral features of religion, as described by scholars like Wald, Silverman, and Fridy.
Consistency with Findings: This politicization explains why Ivorian respondents placed a greater emphasis on religion as a social identity, were less willing to marry across religious lines, and felt closer to co-religionists.
Speed of Politicization Evidence:
A survey question asked respondents about changes in prayer frequency over the past 10 years (relative to 2005 fieldwork).
Ivorian respondents were more than 4.5 times more likely to say they pray 'more' than 'less' (55\% vs. 12\%) compared to Burkinabé respondents who were roughly 2.5 times more likely (42\% vs. 17\%$) (p=0.10 by cluster mean).
This suggests Ivorians became more religious faster, aligning with the differential politicization of religion during the early years of the Ivorian conflict.
Burkina Faso Counterpoint (National Identity):
Burkinabé respondents emphasized nationality (42\%) as their most important identity, a high figure compared to the sub-region.
This could be due to the dominant size of the Mossi ethnic group, cross-cutting cleavages, and a gradual, non-antagonistic ascendancy of Islam and Christianity.
Some argue that the prominence of national identity in Burkina Faso is a secondary effect of the civil strife in Côte d’Ivoire, where the absence of conflict accentuated attachments to nationality, supporting the idea that the CI conflict drove identity changes on both sides of the border.
Conclusion
Main Finding: Living on the Côte d’Ivoire side of the arbitrary border with Burkina Faso is significantly associated with:
A higher likelihood of ranking religion as one's most important social identity.
Favoring co-religionists over co-nationals.
Less willingness to marry someone of a different religion.
Causal Explanation: The most likely explanation is political, rooted in the greater exposure of Ivorian border residents to the politicization of religion during the Ivorian civil war.
Consistency with Theory: These findings are consistent with an instrumentalist political account of religious identity salience and demonstrate how elite mobilization efforts can affect the priority people attach to social identities.
Broader Implications: Many recent conflicts, seemingly religious, have non-religious origins (e.g., Central African Republic, Congo–Uganda border, Israeli–Palestinian conflict). It is crucial to examine political factors before attributing religious salience solely to social, economic, or religious contexts.
Additional Contributions:
Speed of Identity Salience Change: Provides a non-electoral example of rapid shifts in identity salience, occurring within a few years (e.g., 2000-2005 during the Ivorian conflict).
Interplay of Identity Dimensions: Shows how conflicts rooted in one dimension (ethno-national discrimination, access to political opportunities) can affect the salience of other identity dimensions (religious identities), especially due to
This study investigates when religion becomes a significant social identity, proposing that political events and elite mobilization can alter its importance. Utilizing a quasi-experimental design based on the 19th-century partition of Africa—specifically, the arbitrary border between Burkina Faso and Côte d’Ivoire—the research minimizes other causal variables by focusing on populations with similar social, economic, cultural, and historical backgrounds.
Survey data from border regions revealed striking differences: Ivorian respondents were significantly more likely to prioritize religion as their most important identity, less willing to marry across religious lines, and favored co-religionists over co-nationals, compared to their Burkinabé counterparts. These differences persisted even after controlling for potential biases like migration.
The study systematically ruled out alternative explanations, including religious demographics, wealth and secularization, and the quantity or quality of religious institutions, as primary drivers of this disparity.
The preferred explanation centers on the political mobilization of religion in Côte d’Ivoire, particularly after 1993. The post-Houphouët-Boigny era saw politicians strategically exploiting ethnic and religious differences through policies like Ivoirité, which escalated into the Ivorian civil war. This politicization, where religious identity became intertwined with political conflict, strongly aligns with the observed increase in religious salience among Ivorians.
In conclusion, the findings demonstrate that living on the Côte d’Ivoire side of the border is associated with a higher salience of religious identity, primarily due to the political exploitation of religion during the Ivorian conflict. This underscores how elite mobilization efforts can rapidly shift the importance individuals attach to their social identities, suggesting that many seemingly religious conflicts have non-religious, political origins.