Religious fragmentation, social identity and cooperation: Evidence from an artefactual field experiment in India
Introduction
Economists have established over the last two decades a negative relationship between social fragmentation (typically defined as a function of the relative size of different social groups in the population) and economic performance, in particular public good provision (Easterly and Levine, 1999, Alesina et al., 1999, Alesina et al., 2003, Alesina and La Ferrara, 2005, Banerjee et al., 2005). There are many possible reasons for this negative relationship: different social groups may prefer different types of public goods (Poterba, 1998); restricting economic transactions to within a group may also be useful as it reduces informational asymmetries, and increases the scope for the punishment of transgressors (Greif, 1993, Miguel and Gugerty, 2005). Finally, individuals may prefer to share a public good with those of their own group and/or dislike sharing a public good with people from other social groups (Alesina and La Ferrara, 2005).
We study the effect of village-level religious fragmentation on intra- and inter-group cooperation. We are particularly interested in understanding the role religious identity plays in explaining behavior in two classic cooperation games. These games feature incentive structures relevant to many problems pertaining to development contexts. India is an ideal setting to study our research question, having a longstanding social structure characterized by fragmentation along religious lines, as well as a rigid caste system. Issues such as social exclusion and public good provision (or lack thereof) along religious lines are widely documented (de Hann, 1997, Sen, 2000, Bardhan et al., 2010, Das et al., 2011).1 We build upon existing household survey work on religious-based social exclusion in villages in West Bengal, India, and we focus our attention to the problem of religious discrimination among Muslim and Hindu communities in West Bengal.
When reviewing the literature on the effects of ethnic diversity on economic outcomes, Alesina and La Ferrara (2005) identify social identity theory as a means to provide micro-foundations for theoretical explanations as to why fragmentation affects economic performance. Social identity theory argues individuals attach utility to group membership and to the wellbeing of fellow group members to the detriment of outsiders – see Akerlof and Kranton (2000) and Basu (2007) for theoretical analyses of how social identity can affect economic decisions. This paper presents data from an artefactual field experiment examining how social fragmentation can impact cooperation through social identity.
Social identity theory argues that membership of a social group means individuals display a higher concern for the welfare of fellow “in-group” members than outsiders (Tajfel et al., 1971). Experimental evidence from the lab and field supports this inter-group discrimination hypothesis in dictator games (Bernhard et al., 2006, Chen and Li, 2009), distribution games (Klor and Shayo, 2010), and prisoners׳ dilemma games (Goette et al., 2006, Charness et al., 2007), using both artificial and real social groups. Therefore, individuals should be more likely to cooperate with in-group members, even if doing so incurs them an economic cost, and particularly so if total welfare within the group increases as a result. Conversely, individuals may be less likely to cooperate with outsiders. The net effect is therefore a decline in cooperation as the number of outsiders increases (Smith, 2011).
Alternatively, individuals may only be willing to cooperate on a common enterprise if they believe others are likely to do so as well. In that sense, cooperation can take the guise of a coordination problem. The absence of formal institutions that facilitate coordination could lead to economies not achieving desirable equilibria (see Cooper, 1999 for a review of theory and experimental evidence on coordination games with applications to macroeconomics). While coordination problems are a feature of any economy, they are a crucial issue in many developing countries. In these economies, either property rights are not institutionally assured, or access to legal recourse in case of a dispute may be limited and/or costly (see Posner, 1998 and references therein). In this context, the belief by an economic agent about her counterpart׳s willingness to abide by an informal agreement is essential for economic activity to take place. In this context, a sense of group identity could help cooperation to the extent that individuals believe fellow in-group members are more likely to cooperate than outsiders (Brewer et al., 1986, Yamagishi and Kiyonari, 2000). Social psychological evidence suggests these expectations are stronger within group boundaries than across group divides (Tanis and Postmes, 2005; Yamagishi et al., 1999).
To study the effect of social fragmentation on cooperation, we consider two classic cooperation games. The first is the prisoners׳ dilemma, in which cooperation is a strictly dominated strategy for individuals who care only about their own monetary payoff. Either player can unilaterally increase total welfare (at a personal cost) by cooperating. Experimental economists and social psychologists have looked at cooperation in the prisoners׳ dilemma as a measure of other-regarding preferences: if players care sufficiently for the welfare of their counterpart (or about the sum of payoffs), then it can be a dominant strategy to cooperate. However, if both players display other-regarding preferences, then the prisoners׳ dilemma in money payoffs can turn into a coordination game in utility payoffs. In that case, the outcome in which both players cooperate is still an equilibrium, but only if both players believe their counterpart will cooperate.2
The second game is the stag hunt game, in which two players must decide whether or not to cooperate. While defecting ensures a positive payoff, cooperating only pays off if the other player cooperates as well; otherwise the payoff from cooperation is zero. This means both players cooperating can also be an equilibrium of the game, provided players assign high enough probability to their counterpart doing so. Importantly, and unlike the prisoners׳ dilemma, a player׳s own other-regarding preferences play no role in determining behavior in this game, but beliefs about the other player׳s action do.
We study the effect of religious identity among Hindu and Muslim groups by varying the way our subjects are matched with each other. We implement in-group/in-group treatments where Muslim subjects play with fellow Muslim subjects and Hindu subjects play with fellow Hindu subjects; we also implement in-group/out-group treatments where Hindu subjects play with Muslim subjects. Finally, we have a control treatment where the identity of a subject׳s match is uncertain. To study the effect of fragmentation, we resort to a quasi-experimental approach. We take religious composition of villages as fixed, based on the village-level survey on religious fragmentation by Das et al. (2011). We select villages in two districts in West Bengal which conform to one of three categories: Muslim-dominated, where over 90% of the population is Muslim; Hindu-dominated, where over 90% of the population is Hindu; and fragmented, where the Muslim and Hindu communities are roughly equal.3 Our experimental design combines identity treatments with village types to understand how social identity interacts with fragmentation.
We find evidence of in-group favoritism in fragmented villages in the prisoners׳ dilemma, in that cooperation rates are higher in in-group matches than in both unknown and out-group matches. We find a very similar pattern of results in the stag hunt game, although it is of a smaller magnitude and it is not statistically significant. We find no evidence of out-group prejudice (defined as lower cooperation with an out-group member than with an unknown individual) in both the prisoners׳ dilemma and stag hunt games. In both games, cooperation rates among individuals of the same religion are lower in homogeneous villages than fragmented villages, even when controlling for village characteristics such as size, unemployment and literacy rates, which could be proxies for social norms that can predict cooperation. Furthermore, we find cooperation rates amongst individuals of the same religion in homogeneous villages are no different to cooperation rates between two individuals of different religions in fragmented villages.
The rest of the paper is organized as follows. Section 2 gives a brief background to Hindu–Muslim relations which motivate our study. Section 3 describes the experimental design and procedures, Section 4 outlines the results and Section 6 discusses the results and concludes the paper.
Section snippets
Background to Hindu–Muslim relations in India
Studying social identity is complex, especially given the difficulty in isolating the different identities which play a role in social and economic contexts. A key social institution which forges an individual׳s identity in India is religion. Recent Indian history has witnessed several episodes which have stoked traditionally tense relations between Hindus and Muslims. The partition on Bengal along Hindu–Muslim lines in 1905 and the second partition of Bengal into West Bengal and East Pakistan
The games
We report data from two games, each of which captures important features of social behavior: the prisoners׳ dilemma (PD) game and the stag hunt (SH) game. We now briefly describe each game as it was implemented in our experiments, and our approach in analyzing and interpreting behavior in each of the games.
The PD game is the quintessential social dilemma, in which private incentives run against the welfare of the group. It is one of the most widely studied games by scholars investigating the
Results
We start by examining behavior in the PD game, followed by the analysis of the data from the SH game. We complete the analysis by combining our behavioral data with post-experimental survey data which includes measures of attitudes towards out-group members and socio-economic characteristics at both individual and village level. In our analysis, we use each individual׳s decision as an independent observation. We first test for treatment effects using standard statistical techniques and we
Discussion
We now discuss some of the results from our experiment, and we highlight some limitations in interpreting the data. Our first main result is that in both prisoners׳ dilemma and stag hunt games, cooperation rates in homogeneous villages among individuals of the same religion are no different than cooperation rates in our control treatment where the identity of matches is uncertain, which was conducted in fragmented villages. We relied on a quasi-experimental approach to measure the effect of
Conclusion
Social identity theory has been identified as a potential cause for the negative correlation between the degree of social fragmentation and economic performance, particularly the provision of public goods (Alesina and La Ferrara, 2000, Alesina and La Ferrara, 2005). The argument is that individuals derive benefit from membership of social groups and display favoritism towards members of their group at the potential detriment of outsiders. As societies become more fragmented, discrimination
Acknowledgments
We gratefully acknowledge financial support from the ESRC Grant ES/J018643/1. We thank Jharna Phanda, Sujit Kumar Mondal, Ramkrishna Mondal, Naresh Chandra Mondal, Sk.Md. Basiruddin for their excellent assistance with running the sessions. We thank two anonymous referees and the editors, in addition to Avner Ben-Ner, Martin Leroch, Pushkar Maitra, Kim Peters, Rohini Somanathan, Daniel Zizzo, participants at the 2015 SAET conference at U. of Cambridge, the 2014 THEEM workshop at the U. of
References (65)
- et al.
The effect of social fragmentation on public good provisionan experimental study
J. Behav. Exp. Econ.
(2014) - et al.
Managing diversity by creating team identity
J. Econ. Behav. Organ.
(2005) - et al.
Group size and the voluntary provision of public goodsexperimental evidence using large groups
J. Public Econ.
(1994) - et al.
Social identity and preferences over redistribution
J. Public Econ.
(2010) - et al.
Ethnic diversity, social sanctions, and public goods in Kenya
J. Public Econ.
(2005) - et al.
Group identity and relation-specific investmentan experimental investigation
Eur. Econ. Rev.
(2013) - et al.
Enforcement of contribution norms in public good games with heterogeneous populations
Games Econ. Behav.
(2013) Group composition and conditional cooperation
J. Socio-Econ.
(2011)Riot After Riot: Reports on Caste and Communal Violence in India
(2003)- et al.
Economics and identity
Q. J. Econ.
(2000)
Public goods and ethnic divisions
Q. J. Econ.
Fractionalization
J. Econ. Growth
Participation in heterogeneous communities
Q. J. Econ.
Ethnic diversity and economic performance
J. Econ. Lit.
The Nature of Prejudice
Caste as an impediment to trade
Am. Econ. J.: Appl. Econ.
Impure altruism and donations to public goodsa theory of warm-glow giving
Econ. J.
History, social divisions, and public goods in rural India
J. Eur. Econ. Assoc.
Impact of political reservations in West Bengal local governments on anti-poverty targeting
J. Glob. Dev.
Group affiliation and altruistic norm enforcement
Am. Econ. Rev.
An analysis of rural-to-rural migration in India
J. Int. Dev.
The Production of Hindu–Muslim Violence in Contemporary India
Ethnocentrism and its role in inter temporal trust
The social selfon being the same and different at the same time
Personal. Soc. Psychol. Bull.
The psychology of prejudiceingroup love or outgroup hate?
J. Soc. Issues
Understanding social preferences with simple tests
Q. J. Econ.
Individual behavior and group membership
Am. Econ. Rev.
The potential of social identity for equilibrium selection
Am. Econ. Rev.
Group identity and social preferences
Am. Econ. Rev.
Cited by (22)
Religious identity, trust, reciprocity, and prosociality: Theory and evidence
2024, Journal of Development EconomicsDoes ethnic heterogeneity decrease workers’ effort in the presence of income redistribution? An experimental analysis
2023, European Journal of Political EconomyShould regulators always be transparent? a bank run experiment
2021, European Economic ReviewReligious identity, between-group effects and prosocial behavior: Evidence from a field experiment in China
2021, Journal of Behavioral and Experimental EconomicsCitation Excerpt :As investors, both Hindus and Muslims showed ingroup favoritism; however, as investees, they did not display ingroup favoritism. S. Chakravarty, Fonseca, Ghosh and Marjit (2016) studied the role of religious fragmentation in intra- and intergroup cooperation through prisoners’ dilemma game in India. They conducted ingroup/ingroup treatments in which Muslim subjects and Hindu subjects played with subjects similar to themselves; they also implemented ingroup/outgroup treatments in which Hindu subjects played with Muslim subjects.
Prior interaction, identity, and cooperation in the Inter-group Prisoner's Dilemma
2019, Journal of Economic Behavior and OrganizationReligious fragmentation, social identity and other-regarding preferences: Evidence from an artefactual field experiment in India
2019, Journal of Behavioral and Experimental EconomicsCitation Excerpt :There is a small literature on religious identity and cooperation to which our results may speak. Chakravarty et al. (2016b) study the role of religious affiliation and religious fragmentation along Hindu-Muslim lines on cooperation in the Prisoners’ Dilemma and Stag Hunt game in West Bengal, India. They vary the religious identity of the two players, as well as the degree of religious fragmentation at the village level (through choice of village).