|
Activity Number:
|
27
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #306508 |
|
Title:
|
Estimation and Inference in Parametric Stochastic Frontier Models: a SAS/IML Procedure for a Maximum Likelihood Bootstrap Method
|
|
Author(s):
|
Sylvie Tchumtchoua*+
|
|
Companies:
|
University of Connecticut
|
|
Address:
|
210 Quinebaug, Storrs, CT, 06269,
|
|
Keywords:
|
linear model ; bootstrap ; technical efficiency ; inference ; maximum likelihood ; SAS/IML
|
|
Abstract:
|
Parametric Stochastic Frontier Models (PSFM) specify the output or cost of a production unit in terms of a response function and a composite error made of a symmetric noise and a one-sided error representing technical inefficiency. PSFM are widely used in productivity analysis and are commonly estimated using FRONTIER or LIMDEP packages which do not provide inference about the inefficiency term. Moreover usual approach for inference about efficiency in PSFM is based on percentiles of the estimated distribution of the one-sided error term, conditional on the composite error, rather than on the sampling distribution of the inefficiency estimator. We propose a concise program for a maximum likelihood bootstrap method which makes inference about the inefficiency term based on its sampling distribution. The program is written using matrix language SAS/IML with the optimization subroutine NLPQN.
|
