Abstract In a recent article, J.A. Kay proposed a useful measure of the deadweight loss arising from a commodity tax system. The measure answers the question: How much more would the taxed consumer be willing to pay in a lump sum rather than as a commodity tax? Kay's computation of the marginal deadweight loss does not yield the change in this measure for small changes in commodity tax rates, however. This note clarifies Kay's otherwise excellent contribution, derives the measure for Cobb–Douglas utilities, and examines a useful property of it.