横截面与时间序列的相关异质——再论面板数据模型及其固定效应估计

摘要

The paper intends to investigate and explain the two dimensional nature of panel data and its essence of estimation with varieties of fixed effects setting. Conclusions are shown that. cross-section fixed effects estimation is actually a time series regression with a linear constraint; time fixed effects estimation is actually a cross-section regression with linear constraint. Moreover, bilateral effects estimator is the weighted average of pool, cross-section fixed and time fixed effects estimator; Thus, due to its two dimension characteristic, the correlation between cross-section dimension and time series dimension may be significantly different, and the correlation discrepancy could be an important cause for the significant difference between estimation results of different effect settings.%本文的目的在于探讨和解释面板数据的二维特性及其不同的固定效应估计的本质。研究表明，截面固定效应实际上是有线性约束的时间序列回归，其回归系数体现了样本在时序维度的相关关系；时间固定效应实际是有线性约束的横截面回归，其回归系体现了样本在横截面维度的相关关系；双向固定效应估计量则是混合效应、截面固定效应和时间固定效应估计量的加权平均；由于面板数据的二维特性，变量在时序维度的相关关系和横截面维度的相关关系有可能是异质的，这种异质性是导致不同效应设定导致估计结果有很大变化的重要原因。