这两天在学习随机前沿分析。
用模型1测算的时候,得出如下结果:
the final mle estimates are :
coefficient standard-error t-ratio
beta 0 0.98714652E+01 0.26995938E+00 0.36566484E+02
beta 1 0.36796355E+00 0.17642433E-01 0.20856735E+02
beta 2 0.51728147E+00 0.20200918E-01 0.25606830E+02
sigma-squared 0.88942052E+00 0.14104049E+00 0.63061362E+01
gamma 0.94956630E+00 0.88472258E-02 0.10732927E+03
mu is restricted to be zero
eta 0.11997745E+00 0.56855190E-02 0.21102286E+02
log likelihood function = -0.29149648E+03
LR test of the one-sided error = 0.17594617E+04
with number of restrictions = 2
[note that this statistic has a mixed chi-square distribution]
但是当我加入Zit的变量运用模型2进行技术无效函数估计的结果,单随机前沿函数就差别很大(这个是必然的么?)另外,我想问可否把模型1中的技术效率拿出来,单独对Zit进行回归分析呢?
请指教,谢谢。