Capacity is defined as that output which can be produced at minimum average total cost, given the existing stock of plant and equipment and existing techniques and factor prices. The level of capacity is inferred from observed investment behavior. Regression methods are used to estimate a relationship between desired capital stock and several explanatory variables including output, relative prices and time, on the hypothesis that net investment occurs in proportion to the excess of desired over actual stock. The relationship between desired capital stock and output is then inverted to yield a corresponding relationship between capacity and actual capital stock for given prices and techniques. The method is used to calculate aggregate capacity annually for 1949–60 and the properties of the resulting estimates are discussed. New estimates of capacity and its utilization in manufacturing are also presented and compared with those of other investigators.