In a public good environment with positively correlated types, we character
ize optimal mechanisms when agents have private information and can enter c
ollusive agreements. First, we prove a weak-collusion-proof principle accor
ding to which there is no restriction for the principal in offering weak-co
llusion-proof mechanisms. Second, with this principle, we characterize the
set of allocations that satisfy individual and coalitional incentive constr
aints. The optimal weakly collusion-proof mechanism calls for distortions a
way from first-best efficiency obtained without collusion. Allowing collusi
on restores continuity between the correlated and the uncorrelated environm
ents. When the correlation becomes almost perfect, first-best efficiency is
approached. Finally, the optimal collusion-proof mechanism is strongly rat
ifiable.