Test for randomness of the technology parameter in a stochastic frontier regression model

Abstract

Random coefficient models, Stochastic frontier models, Technical Efficiency, Test for randomness,

Rank Test for Testing Randomness of the Technology Parameters in a Stochastic Frontier Regression Model

“…For finding critical points of rank test we used the bootstrap distribution with B = 1 000 and n = 500. We compare the performance of the rank test to the approximate score test considered in Ramanathan and Ghadge (2010) for the same problem. As indicated in that article, we continue to use the same test in the case when u i are exponential.…”

“…A parametric test will be ideal for such a testing problem, when we are absolutely sure about the distributions of the measurement error term and the inefficiency factor in the model. In a recent article, Ramanathan and Ghadge (2010) suggested an approximate score test for this problem assuming half normal distribution to the inefficiency factor and normal distribution for the measurement error. The parametric procedure breaks down when the assumed distributions deviate from the actual one.…”