The persistence of cooperation in public-goods experiments has become an important puzzle for economists. This paper presents the first systematic attempt to separate the hypothesis that cooperation is due to kindness, altruism, or warm-glow from the hypothesis that cooperation is simply the result of errors or confusion. The experiment reveals that on average about half of all cooperation comes from subjects who understand free-riding but choose to cooperate out of some form of kindness. This suggests that the focus on errors and "learning" in experimental research should shift to include studies of preferences for cooperation as well.

This paper presents an experimental test of the proposition that government contributions to public goods, funded by lump-sum taxation, will completely crowd out voluntary contributions. It is found that crowding-out is incomplete and that subjects who are taxed are significantly more cooperative. This is true even though the tax does not affect the Nash equilibrium prediction. This result is taken as evidence for alternative models that assume people experience some private benefit from contributing to public goods.

Models of giving have often been based on altruism. Examples include charity and intergenerational transfers. The literatures on both subjects have centered around neutrality hypotheses: charity is subject to complete crowding out, while intergenerational transfers are subject to Ricardian equivalence. This paper formally develops a model of giving in which altruism is not "pure." In particular, people are assumed to get a "warm glow" from giving. Contrary to the previous literature, this model generates identifiable comparative statics results that show that crowding out of charity is incomplete and that government debt will have Keynesian effects.

The persistence of cooperation in public-goods experiments has become an important puzzle for economists. This paper presents the first systematic attempt to separate the hypothesis that cooperation is due to kindness, altruism, or warm-glow from the hypothesis that cooperation is simply the result of errors or confusion. The experiment reveals that, on average, about half of all cooperation comes from subjects who understand free-riding but choose to cooperate out of some form of kindness. This suggests that the focus on errors and 'learning' in experimental research should shift to include studies of preferences for cooperation as well.

This paper presents an experimental test of the proposition that government contributions to public goods, funded by lump-sum taxation, will completely crowd out voluntary contributions. It is found that crowding-out is incomplete and that subjects who are taxed are significantly more cooperative. This is true even though the tax does not affect the Nash equilibrium prediction. This result is taken as evidence for alternative models that assume people experience some private benefit from contributing to public goods.

Private providers of public goods, such as charities, invariably enlist fund‐raisers to organize and collect contributions. Common in charitable fund‐raising is seed money, either from a government grant or from a group of “leadership givers,” that launches the fund drive and generates additional gifts. This paper provides a theoretical basis for fund‐raisers and seeds to charity. The primary assumption is that there is a range of increasing returns at low levels of provision of the public good. It is shown that fund‐raisers have a natural and important role, and that sometimes only a small amount of seed money can grow into a substantial charity.

Models of giving have often been based on altruism. Examples include charity and intergenerational transfers. The literatures on both subjects have centered around neutrality hypotheses: charity is subject to complete crowding out, while intergenerational transfers are subject to Ricardian equivalence. This paper formally develops a model of giving in which altruism is not "pure." In particular, people are assumed to get a "warm glow" from giving. Contrary to the previous literature, this model generates identifiable comparative statics results that show that crowding out of charity is incomplete and that government debt will have Keynesian effects.

Can the well-known experimental phenomenon of present-bias in intertemporal choice be confounded with the risks associated with receiving payment? Can measurements of risk preferences be used to represent desires for smoothness in intertemporal payments? In our two 2012 papers in this journal we explored these two questions and found the answer to the first to be yes and the second to be no. We feel the three papers inspired by our work and published here underscore these arguments and point to interesting new possibilities for modeling and measuring risk over time.