The optimal pollution permit trading system is examined when the regulator, faced with incomplete information on firms' abatement costs and delivery coefficients, seeks to minimize expected total abatement costs to meet an ex ante pollution target. Intuitively, we find that the optimal trading ratio and permit cap are set such that there will be more pollution when abatement costs are high and less pollution when abatement costs are low. Surprisingly however , even when the delivery coefficients are known with certainty, the optimal trading ratio will not necessarily equal the delivery coefficient, nor will it be optimal for the total permit quantity to equal the given pollution target. Instead, the trading ratio will tend to be larger when there is negative correlation between firms' abatement costs and/or there is positive correlation between abatement costs and the delivery coefficient. The result that the optimal permit trading system under these circumstances depends on the regulator's information on firm's abatement costs contrasts sharply with a previously well-established attribute of permit trading systems: no information on firms' abatement costs is needed in order to design an optimal policy and achieve the least cost of reaching a pollution target (ex post). Our results demonstrate that whether an ex ante or ex post pollution target is used has fundamental implications for the design of the permit trading system. Finally, while not descriptive of all pollution problems, the class of pollutants for which this model applies is large.