This paper makes several contributions. First, we demonstrate the generality of the Fullerton, Karney, and Baylis (2012) model by showing cases where leakage can exceed 100 percent. We solve for conditions under which total emissions increase or decrease. We also solve for welfare effects, and for “cost effectiveness” (the additional welfare cost per ton of net abatement). And we explore the relationship between the sign of leakage and the sign of the effect on welfare.