非参数工具变量法及应用

发布时间：2018-06-21 20:40

本文选题：计量经济学 + 工具变量法　；参考：《东北师范大学》2017年硕士论文

【摘要】：工具变量法是计量经济学中一种重要的方法,用于解决解释变量的内生性问题。工具变量法需要寻找与内生解释变量相关,又与误差无关的变量,即工具变量,并在此基础上解决问题。这种方法在理论上和应用上都具有很大的价值,在理论上建立了一整套方法,在经济,教育,医学等领域中,解决了许多实际问题。目前在实际应用中,以参数工具变量法为主,需要事先设定模型。但是解释变量和被解释变量之间的关系往往很复杂,之间的确切关系我们并不知道。这时候,我们需要一种方法来近似这种关系,非参数工具变量法就是其中的一种,它用一组标准正交基来近似解释变量与被解释变量之间的关系,得到的估计具有简单的表达形式,易于计算,并具有良好的大样本性质。本文用对数日均乘客量与对数票价之间的关系的实证研究揭示了参数工具变量法与非参数工具变量法之间的相同点和不同点,首先发现对数票价是一个内生解释变量,然后寻找工具变量,即出发地航空公司所占的最大份额,在使用简单最小二乘法和两阶段最小二乘法时,存在需要控制对数距离和对数距离的平方的问题,由于对数距离和对数票价是相关的,这暗示着对数日均乘客量与对数票价之间存在非线性关系,最后使用非参数工具变量法得到了对数日均乘客量与对数票价之间的非线性关系。本文最后对非参数工具变量法的前景进行了展望。由于非参数工具变量法具有样本容量越大则模型拟合越精确的特点,在大数据技术发达的今天,这是一种潜力无限的方法。
[Abstract]:Tool variable method is an important method in econometrics, which is used to explain the endogenous problem of variables. The tool variable method needs to find the variable which is related to the endogenous explanatory variable and which is independent of the error, that is, the tool variable, and solve the problem on this basis. This method has great value in theory and application. It has established a whole set of methods in theory, and solved many practical problems in the fields of economy, education, medicine and so on. At present, the parametric tool variable method is the main method in practical application, and the model needs to be set in advance. But the relationship between explanatory variables and interpreted variables is often complex, and the exact relationship between them is unknown. At this point, we need a way to approximate this relationship, and the nonparametric tool variable method is one of them, and it uses a set of standard orthogonal bases to approximate the relationship between the variable and the variable being explained. The obtained estimates have simple expression, easy to calculate and good properties of large samples. An empirical study of the relationship between logarithmic daily passenger volume and logarithmic ticket price reveals the similarities and differences between the parametric tool variable method and the non-parametric tool variable method. Firstly, it is found that the logarithmic ticket price is an endogenous explanatory variable. Then we look for tool variables, that is, the largest share of the departure airline, and when we use the simple least square method and the two-stage least square method, there is the problem of controlling the logarithmic distance and the square of the logarithmic distance. Since logarithmic distance and logarithmic fare are related, this implies that there is a nonlinear relationship between logarithmic daily passenger volume and logarithmic fare. Finally, the nonlinear relationship between logarithmic daily passenger volume and logarithmic ticket price is obtained by using nonparametric tool variable method. Finally, the prospect of nonparametric tool variable method is prospected. Because the non-parametric tool variable method has the characteristics that the larger the sample size, the more accurate the model fitting is, in the modern big data technology, this is a method with unlimited potential.
【学位授予单位】：东北师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：C815

【相似文献】