计量模型中的加总偏误与内生性：一种数值模拟方法

理论文献已经证明了细分数据模型的内生性和加总过程产生的内生性是加总偏误的根本原因。但是由于内生性涉及误差项与回归量之间的相关性问题，试图通过实证方法审视这两类内生性对加总偏误的影响变得比较困难，而数值模拟却是一个较为理想的方法。在数值模拟中，我们通过控制随机变量的分布形式以及随机变量之间的相关程度，进而对两类内生性因素产生的加总偏误进行全面细致的考察。本文的研究将为加总偏误的内生性解释提供有力证据。

A Numerical Simulation Analysis on Aggregation Bias and Endogeneity in Econometric Model

The endogeneity in disaggregation model and in the process of aggregation are proved to be the reasons of aggregation bias in econometric models. However, it is difficult to inspect the relationship between the endogeneity and aggregation bias, because the error term is unknown. Fortunately, numerical simulation is the ideal method to deal with the problem. By controlling the distribution of random variables, we successively inspect the aggregation unbias, aggregation bias and the relationship between endogeneity and aggregation bias. In result, the research strongly supports the opinion of endogeneity and aggregation bias.