This article may be read as a comment on and extension to the recent survey of studies of the social costs of smoking in this journal by Markandya & Pearce (1989), covering some aspects of the literature not included in that survey.

Inevitably, there is a considerable amount of overlap with the material in several of the listed references, including the Markandya & Pearce (1989) survey.

The article argues that the published studies based on the incidence approach to the costs of smoking, even on their own terms and accepting the stated assumptions, are erroneous, yielding cost estimates that are too high.