弱外生性条件下双固定效应模型估计方法的优化

固定效应模型是排除其他干扰因素以确定因果机制的有效工具，在企业面板数据分析等领域得到广泛应用。由于企业个体异质性等原因，普通固定效应方法无法解决估计的冗余参数问题。本文沿着折刀法的思路逐步修正低阶偏差和高阶偏差，在弱外生性条件下构建Half-Panel Jackknife方法，并将适用范围拓展到短面板线性模型，弥补了现有方法的不足。最后，采用企业并购重组和政府补助影响企业创新的两个实例，研究发现，对半面板折刀法在双固定效应模型中能较好地控制企业数据的抽样特征与个体异质性。

Optimization of Double Fixed Effect Model Estimation under Weak Exogeneity

Fixed effect model is an effective tool to exclude other interference factors to determine causal mechanisms, and is widely used in enterprise panel data analysis. Because of individual heterogeneity, the common fixed effect method cannot solve the incidental-parameter problem. According to the idea of Jackknife, the low order and high order deviation are gradually corrected, and the Half-Panel Jackknife method is constructed under the condition of weak externality. The application range is extended to short panel and linear model, which makes up for the shortcomings of the existing methods. Finally, taking M&A and government subsidy as examples, the results show that the Half-Panel Jackknife method can better control the sampling characteristics and individual heterogeneity of enterprise data in the double fixed effect model.