Endowment heterogeneity, incomplete information & institutional choice in public good experiments

•We vary of the observability of heterogenous endowments in a public good experiment.•Groups get experience with central authority and peer punishment before voting which one to implement.•In Unobserved subjects with lower (higher) endowments prefer central authority (peer punishment).•Cooperating low endowment members receive more punishment in Unobserved.•Subjects with higher endowments in Unobserved undermine targeted sanctions by disguising contributions.