A Frontier Approach to Testing the Averch-Johnson Hypothesis

The mathematical programming technique Data Envelopment Analysis is used to test the famous hypothesis of Averch and Johnson that utility regulation leads to overuse of capital because the regulated firm earns a return s greater than its cost of capital r, an implicit capital subsidy resulting in allocative inefficiency. Technical and allocative inefficiency are based on cost and production frontiers from 337 electric generating plants using 1970 data, and r is based on the Capital Asset Pricing Model. Significant capital overuse and general failure to minimize costs is detected, but a second-step regression analysis finds no relationship between the overuse and the s-r subsidy. A small updated data set covering the period 1980-2004 suggests that overuse of capital is no longer a problem, a result that may be owing to recente deregulation and restructuring.