Productivity Analysis of U.S. Manufacturing Using a Stochastic-Coefficients Production Function.

In this article, we show that the production, aggregation, and estimation theories that lead to a stochastic-coefficients production function are more general than those leading to a fixed-coefficients production function. Estimates of productivity changes implied by these two types of Cobb-Douglas production functions are compared for several U.S. manufacturing sectors for the period 1955-1982. All of the estimates of the stochastic-coefficients Cobb-Douglas function have the right algebraic sign for all industries considered, whereas some of the estimates of its fixed-coefficients counterpart have the wrong sign. [ABSTRACT FROM AUTHOR]