The Nonparametric Estimation Of Panel Data And Its Application

Panel data also known as longitudinal data .A panel or longitudinal data set is one that follows a given sample of individuals over time, and thus provides multiple observations on each individual in the sample. Panel data which combines features of a cross-section data and time series, become an important field of the modern of econometrics. At the same time, Nonparametric estimation has also been attented. In general, the model form of economic variables is unknown or uncertain, but we need to estimate the regression function. In this case, nonparametric model is more appropriate in practical applications.When we build econometric models, some factors are often not considered or observed. At this time, there is often error autocorrelation in the model and the estimates of parameter is not unbiased or effective. In this case, the panel data model is introduced, which can get unbiased estimates by not observing actual values when we missed a number of variables. But in most cases we can not determine the specific form of the model, which often generated model error. So we resort to nonparametric estimate to compensate for these deficiencies.The article elaborates a number of applications of panel data in the the economy. In the second part, we introduced various models of panel data model , covariance estimation and generalized least squares estimation. In the third chapter, In the third chapter,we lead to nonparametric estimation through analyzing lacks of parameter estimation. We estimate panel data model by the way of local linear estimation and elaborate asymptotic normality. Finally, through the empirical example, it easy to find that local linear estimation method offers us a new means to do nonparametric estimation for panel data.