面板模型高阶矩估计的理论研究及其应用

Recently, econometric analysis of panel data has received the great attention. It becomes very difficult to obtain respectively the high order moment of error component for panel data models due to the mixing properties of the errors and the effects. Based on the newest method on high order moment estimation of the errors and random effects in panel data models, we study the estimators and their asymptotic normalities of the skewness and kurtosis for the errors and the effects，which led to some tests for the characters of the distribution of the errors and the individual effects such as skewness and kurtosis. On the basis of these results, we can avoid the misspecification in modelling panel data and enhance the reliability of statistical inference. Moreover, we carry out a small simulation study for illustration. As a real example, we analyze the relationship between higher education labor and economic growth in China. Specifically, we compute the moments up to fourth of the errors and the effects, and then test their symmetric and heavy-tailed properties. The results show the asymmetrics and heavy-tailed properties.