A Note on the Efficient Semiparametric Estimation of Some Exponential Panel Models
研究了条件最大似然估计在某些面板模型中的半参数有效性，发现当存在关于固定效应的完全充分统计量且该统计量不依赖于感兴趣参数时，条件最大似然估计量达到半参数效率界，例如面板泊松回归和面板负二项模型。
This paper investigates the semiparametric efficiency of the conditional maximum likelihood estimation in some panel models. The nonparametric component of the model is the unknown distribution of the fixed effect. For the exponential panel model, there exists a complete sufficient statistic for the fixed effect. When the complete sufficient statistic does not depend on the parameter of interest, the conditional maximum likelihood estimator (CMLE) achieves the semiparametric efficiency bound. In particular, the CMLE is semiparametrically efficient for the panel Poisson regression model and the panel negative binomial model.