关键词: 固定效应模型, 移动分块经验似然, 序列相关

Abstract: This paper proposes a moving blocks empirical likelihood method for panel linear regression models with fixed effects, allowing serial correlations with any form in idiosyncratic errors in the model. Large sample properties are provided. Simulation results show that, whether the idiosyncratic errors structure were identified or not, the proposed method works well in both situations. Furthermore, the proposed method is much more effective than those of Baltagi and Li (1994) and Gon?alves (2011). The proposed method is applied to investigate the relationship between CO2 emission and urbanization. The empirical results show that urbanization has significant effects on CO2 emission and the effects during each stage are different.

Key words: Fixed Effects Model, Moving Blocks Empirical Likelihood, Serial Correlation