We show that the famous neutrality result in the theory of public good contributions (Warr, Kemp, Bergstrom, Blume and Varian) depends crucially on the assumption that agents do not take into account the effect of their public good contribution decisions on the relative price of the private goods. Thus, the scope of applicability of their result is not as large as one might at first think. Our non-neutrality results hold even if all countries are identical in technology, preferences, and endowments.