Do fixed-prize lotteries crowd out intrinsically-motivated public good contributions?

Material incentives to contribute can act as a tool for increasing social efficiency in the voluntary contribution mechanism (VCM) for provision of public goods. Evidence on whether such incentives increase the net fundraised amount is mixed, however. Are the incentives weak relative to their cost or do they crowd out contributions driven by intrinsic motivations? We introduce an experimental method aimed at separating the two hypotheses. The method eliminates the material incentive to contribute for one group member (who is compensated for this) while preserving it for the other members. We identify the crowding-out effect by comparing the contribution of the non-incentivized member with his contribution in the VCM. We apply the method to a mechanism that augments the VCM with a fixed-prize lottery with winning probabilities proportional to individual contributions. We find that even though this mechanism increases contributions relative to the VCM, it also significantly crowds out intrinsically-motivated contributions. In the absence of the crowding-out effect, the lottery prize would more than pay for itself, implying the material incentive is strong enough relative to its cost. But in its presence, the lottery does not pay for itself. The proposed decomposition method can analogously be applied to other similar settings.