Content area
Full Text
1. Introduction
Many previous studies of linear public goods experiments have compared two matching designs: “Partners design” (Partners), which is a repeated game, and “Strangers design” (Strangers), which is a repeated single-shot design. For example, Andreoni (1988) found that Strangers contribute more than Partners. However, subsequent works have found it difficult to replicate this result, with the majority of researchers reporting inconclusive evidence in this regard (see, Andreoni and Croson, 2008). Our study aims to contribute to the body of knowledge on this topic by explaining the mixed evidence in the literature on Partner vs Stranger comparisons.
Croson (2007) conducted linear public goods experiments with Partners in which subjects were asked to decide how much of their endowment to contribute to the public good as well as to estimate the mean contributions made by other group members (called “belief” hereafter). The author used a random-effects regression and found a significant positive relationship between belief and contribution, strongly supporting the idea that contributions stem from conditional cooperation (reciprocity theory)[1]. Similarly, Fischbacher and Gächter (2010) conducted linear public goods experiments with Strangers. Their design also required the subjects to estimate belief, and they observed the same result as Croson (2007). Taken together, the results presented by both Croson (2007) and Fischbacher and Gächter (2010) show that conditional cooperation is an important explanation for voluntary contributions regardless of the matching design employed. Hence, one reason for the difference in contributions by Partners and Strangers could be the varying expectations of contributions by group members. Based on the foregoing, this study focusses on conditional cooperation and investigates if the contribution difference between Partners and Strangers is explained by differences in belief. Additionally, we aim to clarify the extent of contribution because of conditional cooperation.
Our linear public goods experiments have the following three features. First, we asked subjects to estimate the contributions of their counterparts, similar to the methods adopted by both Croson (2007) and Fischbacher and Gächter (2010). Second, a group in our study consisted of two subjects[2]. Croson (2007) found that the “median” contribution of the other players is a better predictor of a subject’s own contribution than either the maximum or the minimum contribution. Moreover, the “maximum” contribution coefficient is significant at the p <...