The effects of a policy measure often reach the consumer only after one or more intermediatesteps, for instance because the measure lowers the cost of an input for an industry producinga consumer good. This paper is concerned with the question how to measure such indirect effectscorrectly under conditions of perfect and imperfect competition. Conventional CBA measures the indirecteffects on consumers as the direct effect on other actors (e.g. the Marshallian consumer'ssurplus of the demand for the input whose price changes). Formal analysis establishes thecorrectness of this approach under perfect competition, provided that the demand curve isappropriately defined. Under less than perfect competition, the indirect effect can differ fromthe direct effect. Under monopoly the indirect effect is always larger than the direct effect.Under monopolistic competition it can be smaller, identical or larger, depending on the details ofthe model specification and on the possibility of entry.