برآورد غیر پارامتری در اثر واردات و صادرات تجارت تا رشد اقتصادی در چین

Based on the panel data of 31 provinces, municipalities and autonomous regions in China from 1997 to 2008, this paper uses the method of nonparametric local linear kernel estimation to study the relations of physical capital stock, labor capital, export trade and import trade with GDP. The research results are summarized as follows: (a) the above mentioned four factors have positive significant impacts to GDP; (b) the pointwise regression elasticity estimates of export trade and import trade with respect to four factors have both types of “U” and inverted “U” shapes.

Since reform and open policy has been taken from 1978, the economic development of China has made great progress. The growth rates of import and export trade have been higher than that of GDP. It is one of the primary factors keeping the national economy growth. In recent years, many scholars have studied the relationship between import and export trade and economic growth by using economic theory and econometric methods. Many of them proved the “exportled growth” hypothesis considered from theoretical and empirical perspectives. The related literature includes Azomahou et al. (2006), Su and Ullah (2007), Henderson, Carrol and Li (2008), Shan and Sun (1998), Liu and Feng (2007), Zhou (2008) and etc.. We find that most of their used methods are parametric methods. Especially, econometric analysis of panel data is always based on the parametric panel model. We know that parametric panel model requires a lot of assumptions which are not easily to be satisfied. However, the assumption conditions for nonparametric panel model are very few; the model is mainly decided by the data of variables and has many advantages. Therefore, we try to use nonparametric method to study the impacts of export and import trade to economic growth in China and hope to find some internal phenomena. As we know that nobody has used such method to do this problem.

Based on the panel data of 31 provinces, municipalities and autonomous regions in China from 1997 to 2008, this paper uses the nonparametric local linear kernel estimation method (Ullah and Roy, 1998) and pointwise regression method to study the relations of physical capital stock, labor capital, export trade and import trade with GDP. Compared with traditional parametric panel data model, the nonparametric panel model is not only overcome the multicollinearity of independence variables but also get the changing trend of export trade and import trade with respect to the four factors. Unlike the parametric panel data model has many restrictions, nonparametric panel data model can reflect the true economic phenomenon much better. Our research results can be summarized as follows: 1. The marginal impacts of physical capital stock, labor capital, export trade and import trade to GDP are positive. The influence order from small to large is import trade, labor capital, physical capital stock and export trade respectively. Therefore, we should try to increase the export trade amount of provinces in China, this will enhance the impacts of export trade to economic growth. 2. There are two types of changing trends of export trade and import trade with respect to the four variables: inverted “U” type and “U” type. The trends of output elasticity of export trade with respect to physical capital stock, labor capital, export trade and import trade are inverted “U” types. The trends of output elasticity of import trade with respect to physical capital stock, labor capital, export trade and import trade are “U” type, “U” type, “U” type and inverted “U” type respectively. The corresponding policies and suggestions for keeping economic growth in China can be adjusted according to these characteristics.