Capacity utilization (CU) is usually defined as the ratio of actual output to the output corresponding to (i) the minimum point on the SRATC curve, (ii) the point of tangency between the LRATC and SRATC curves. In practice, however, CU is often measured as the ratio of actual to the maximum potential output consistent with a given capital stock. This paper demonstrates how to estimate the theoretical measures of CU, and examines the correlation between the three measures of CU, and the McGraw-Hill estimates of CU, using data from a sample of US privately owned electric utilities for 1961-83.