Electric utilities in a number of American states devote significant portions of their resources to demand-side management (DSM) programs designed to reduce their customers＇ electricity consumption. As other jurisdictions consider similar programs, the public policy cost-efficiency criteria for determining how much utilities should pay for DSM remain controversial This is paper develops the appropriate measure of the economic benefits and costs of DSM, using a conventional economic welfare framework, and compares it to the standard cost-effectiveness tests used in most jurisdictions today. The standard tests are found to be incomplete, suffering from three potential biases. Modifications to the standard tests are suggested to address each of the biases. A numerical example is used to illustrate the nature and potential magnitude of the bias in the current tests.