A stochastic non-parametric approach for estimating production function

Estimating approach of production function is divided into two handle: Non-parametric (Deterministic): Data Envelopment Analysis (DEA) Parametric (Stochastic): Stochastic Frontier Analysis (SFA) Target of this paper is to combine the advantages of both approaches in a unified structure. We combine non-parametric frontier of DEA with remainder decomposition approach into two components technical inefficiency and noise error of SFA approach. For non-parametric frontier of DEA only impose free disposability, convexity and some specification of return to scale and for two components technical inefficiency and noise error impose standard SFA assumptions. Resulting function always take a piecewise linear form alike to DEA frontier. A simulated example is used to illustrate the method.

Keywords:

production function , data envelopment analysis (DEA) , stochastic frontier analysis (SFA) , non-parametric regression , frontier estimation