Public-good experiments with large groups

Abstract

Many of real-world public-goods are characterized by a marginal per capita return (MPCR) close to zero and have to be provided by large groups. Up until now, there is almost no evidence on how large groups facing a low MPCR behave in controlled public-good laboratory experiments involving financial incentives. Connecting four experimental laboratories located in four di¤erent German universities via Internet, we are able to run such experiments. In ad-dition to the group size (60 and 100 subjects), we vary the MPCR which is as small as 0:02 or 0:04. Our data reveal a strong MPCR effect, but almost no group-size e¤ect. Our data demonstrates that, even in large groups and for low MPCRs, considerable contributions to public goods can be expected. Interestingly, the contribution patterns observed in large and very small groups are very similar. To the best of our knowledge, this study is the first one that includes large-group laboratory experiments with a small MPCR under conditions comparable to previous small-group standard public-good experiments.