ABSTRACT
This study investigated the effect of group representatives on contribution behaviour in response to group members’ requests in repeated public goods games. Data came from 135 students in 15 groups enrolled in the experiment, grouped in the following treatments: no group representatives (NR, NRG) and group representatives (R1, R2). We also tested initial requests wherein the group members’ initial requests (i.e. NRG, R2) mediated the positive relationship between individual-level contribution preferences and group contributions. We used a fixed-effects GLS regression and IV regression to analyse the effect of group representatives and group members’ requests. The results indicated the following: (a) rotating group representative/group members’ requests was related to group contributions at the session (within-members) and group (between-group) levels; (b) the reactions to members’ requests positively predicted group contributions; (c) subjects in the no group representatives treatment formed their requests by relying more on previous group contributions than subjects in the group representatives treatment, who relied more on the previous group members’ requests; and (d) the initial requests explained long-term contribution levels, which resulted in variations in group contributions observed between the NRG and R2 treatments. Our findings highlight the role of rotating group representatives in stimulating cooperation among members, while group members’ requests impel individuals to make contribution decisions at the group level.
Acknowledgments
I would like to thank Prof. Daniel Friedman, Prof. Songsong Li and anonymous referees for constructive and very useful comments, and Max and Annabel for language editing. We are grateful to the LEEPS lab and Experimental Economics Workshop in UCSC.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 These include religion (Benjamin, Choi, and Fisher Citation2016), political beliefs (Benjamin, Choi, and Strickland Citation2010), risk aversion (Kranton et al. Citation2012), and even social preferences (Chen and Li Citation2009; Fehr and Hoff Citation2011).
2 The study differs from prior research wherein a single leader was chosen to decide on behalf of group members in PGGs (see Güth et al. Citation2007; Van Dijk, Wilke, and Wit Citation2003; Van Vugt et al. Citation2004; Van Vugt and Spisak Citation2008).
3 Because the primary research questions concerned subjects’ responses over time as they learned about group members’ actions. The strangers’ group would interfere with the reaction processes to their actions (Andreoni Citation1988; Andreoni and Croson Citation2008; Chaudhuri, Paichayontvijit, and Smith Citation2017).
4 In this part, all participants played an individual 1-shot public goods game with no feedback, revealing preferences for contributions (Auerswald et al. Citation2016).
5 We described the data for the second 15 rounds in another paper (in preparation) and did not analyse them here.
6 Like Chaudhuri, Paichayontvijit, and Smith (Citation2017), we used seemingly unrelated estimation. It was equivalent to a pooled regression, which uses interaction effects to estimate a round coefficient for each treatment.
7 The Hausman test results led us to reject the null hypothesis that the random effects estimator would be consistent, so we report the fixed-effects estimation here. We also applied the random effects and Tobit estimation, which generated similar results.
8 Because the less subjects contribute to the public goods account, the higher payoff subjects would obtain, the coefficients for Payoff revealed the individual contribution tendencies. We also used random effects estimation, including individual specifications, which provided similar results.