摘要

This paper investigates whether risk preferences inform the decision of how much to put into the public account in the public goods game under the three different frames (the two house money effect frames: the standard and covered-loss frames, as well as the real-loss frame). The main contribution of this paper finds that the covered loss and real loss treatments are statistically equivalent. This assures researchers that just introducing the notion of loss into an experimental treatment without the need for participants to realize a real loss is still a valid experimental instrument. We also find that the house money effect is a better explanation for the difference in contributions between gain and loss framing than loss aversion.