Maximal domain for strategy-proof rules with one public good

In the context of the provision of one pure public good, we study how large a preference domain can be to allow for the existence of strategy-proof rules satisfying the no vetoer condition. This question is qualified by the additional requirement that a domain should include "a minimally rich domain." We first characterize generalized median voter schemes as the unique class of strategy-proof rules on minimally rich domains. Then we establish that the unique maximal domain, including a minimally rich one which allows for the existence of strategy-proof rules satisfying the no vetoer condition, is the domain of convex preferences. Classification Number. D71. (C) 2000 Academic Press.