Programs where firms sell emissions “offsets” to reduce their emissions continue to provide important complements to traditional environmental regulations. However in many cases, particularly with current and prospective climate change policy, they continue to be very controversial. The problem of adverse selection lies at the heart of this controversy, as critics of offset programs continue to produce evidence that these projects are paying firms for actions they would have undertaken anyway, and are not producing “additional” reductions. This paper explores the theoretical sources of non-additional offsets. An important distinction arises between sales that indicate adverse selection and those that reveal information about aggregate emissions levels.