Abstract Compare with the classical stochastic frontier model, the problem a spatial lag stochastic frontier model encountered is how to estimate the model? Because of spatial lag effect, the classical corrected ordinary least estimator is failed and the maximum likelihood method, although available, will encounter computation burden problem. This paper constructs a new corrected ordinary least square estimator of the half normal spatial lag stochastic frontier model, which estimates spatial lag model by ordinary least square estimator and adjusts the estimates of constant and variance by considering stochastic frontier effect. Simulation studies demonstrate that the method we proposed has good sample property.
|
Received: 12 May 2014
Published: 15 August 2014
|
Corresponding Authors:
ZHANG Jinfeng
E-mail: zjf80125@163.com
|
|
|
|