Traditional analysis of public goods like income redistribution explicitly assumes that people's preferences toward a specific good differ. Given this assumption, we see no reason why some people's demands cannot extend downward into the negative quadrant or originate in the negative quadrant. In short, we see reason why a government good or service cannot be a ‘public good’ to some people and, at the same time, a ‘public good’ to other people. The purpose of this note has been to draw out the logical implications of the assumption that people's individual demands are spread out over the first and fourth quadrants. We have shown that such as assumption requires an adjustment in the way the demand for a public good is constructed and that such an assumption requires that we adjust significantly our conclusions regarding the Pareto efficiency of collectivization of aid to the poor. A central conclusion drawn from the analysis is that a move from private to public charity may very well lead to a reduction in the amount of aid received by the poor and may, therefore, be Pareto inefficient. Further, we have concluded that people other than the poor may, because of the public poverty programme, be necessary beneficiaries of the public transfer system. There may be good arguments for public charity; however, the public goods argument is not necessarily one of them. These conclusions should not seem bizarre. When the political process permits income redistribution, it should surprise no one that people with negative as well as positive demands for poverty relief allow their preferences to be known to the people who must act on the legislation. In addition, it should surprise no one to learn that political compromises, involving public subsidy programs for the poor and the not-so-poor, are, from time to time, developed and that the welfare state, for that reason and to that extent, can expand substantially beyond the bounds implied in efforts to help only the truly poor.