Formulating And Estimating Dynamic Stochastic Production Models

Both the dynamic and stochastic aspects of economic life are increasingly being recognized and incorporated in current analytics. In this paper we examine these aspects in production models, with the aim of achieving the following goals: (i) Unifying the various existing efforts into one more general analytic framework; (ii) Evaluating the tractability of the two dominant functional forms in the presence of dynamics and uncertainty; (iii) Exhaustively analyzing the estimation issues of the general form in stochastic multiperiod and multistage problems with observable and unobservable intermediate outputs; (iv) Determining the conditions under which production function and factor demand estimation is separable, i.e., when knowledge of the production process is not required for efficient factor demand estimation, and vice-versa; (v) Examining the existence of analytic factor demands in the general dynamic stochastic model; and finally, (vi) Providing alternative conditions that lead to an assortment of tractable, estimable models for empirical use. We are encouraged by the conclusion that, subject to avoiding certain pitfalls, the general production model can be extended to include both multiple periods and uncertainty, with increased empirical relevance.