基于非参数方法对外汇占款、股票指数的分析研究

发布时间：2018-01-09 01:12

本文关键词：基于非参数方法对外汇占款、股票指数的分析研究　出处：《天津财经大学》2013年硕士论文　论文类型：学位论文

【摘要】：居民消费价格指数(Consumer Price Index,简称CPI)与民众的生活有紧密的联系。CPI可以用来衡量货币的购买力,可以反映工资水平,一定程度上反映通货膨胀的程度。物价稳定是国家宏观经济政策的目标之一,突显其重要性。自次贷危机以来,居民消费价格指数一直呈现不稳定的状态。研究外汇占款、股票市场指数、人民币汇率对CPI的影响,具有重要的理论意义和应用价值。参数回归模型是分析时间序列的一种强大的工具,其回归函数的形式是已知的。如果能够确定总体的分布,此模型的拟合效果很好。但是金融危机发生后,我国经济变量的数据结构有所变化,表现出非线性特征,如非正态性、非对称、双峰性、异方差性或者滞后变量之间存在非线性关系,而且数据之间的参数关系很难明确。采用传统的计量模型就会出现误差。在这种情况下非参数回归模型是一个很好的选择。非参数回归模型对回归函数的形式无限制,对解释变量和被解释变量的分布也很少有约束。与古典回归模型相比较,非参数回归模型有高拟合度的优点,解释能力更强,作出的推断精度更高。因此不再遵循传统的方法,而是首先根据经济学理论重点阐释了外汇占款、股票市场指数、人民币汇率以及CPI之间的相互关系。其次建立向量自回归模型对CPI的影响因素进行分析,结果表明人民币汇率由于受到国际资本流动和商品贸易两种相反的作用,对CPI的影响并不完全,股票市场对CPI的影响并不显著,外汇占款对CPI的影响显著。再次针对外汇占款对于居民消费价格指数的影响运用非参数方法进行更为深入的分析。结果发现在不同时期CPI受外汇占款的影响是不同的,两者并不是完全正相关的关系。最后给出了相关的政策建议。论文的创新之处：第一,对参数回归模型、非参数回归模型、半参数回归模型进行了比较,总结了各自的优缺点及适用范围,方便他人使用。第二,针对外汇占款对于居民消费价格指数的影响运用非参数核回归的方法进行了分析研究。
[Abstract]:The consumer price index (Consumer Price Index, referred to as CPI) is closely related to the.CPI can be used to measure the purchasing power of money and people's life, can reflect the level of wages, to a certain extent reflect the degree of inflation. The price stability is one of the goals of national macroeconomic policies, to highlight its importance. Since the subprime crisis, the consumer the price index has been in an unstable state. The research of foreign exchange, the stock market index, the impact of RMB exchange rate on CPI, which has important theoretical significance and application value.
The parametric regression model is a powerful tool for the analysis of time series, the form of the regression function is known. If we can determine the overall distribution, the fitting effect of this model is very good. But after the financial crisis, the data structure of our country economic variables change, shows nonlinear characteristics such as non normality Shuangfeng, asymmetric, heteroscedasticity or lag nonlinear relationship between the variables and parameters of the relationship between the data is difficult to identify. The traditional measurement model will appear error. In this case the nonparametric regression model is a good choice. No restrictions on non parametric regression model for regression function., have constraints on the explanatory variables and the variables distribution is very small. Compared with the classical regression model, nonparametric regression model has high fitting advantages, stronger, inference accuracy More high. Therefore no longer follow the traditional method, but according to the theory of economics explains the focus of foreign exchange, the stock market index, the relationship between RMB exchange rate and CPI. Secondly analysis of factors influencing the regression model of the CPI vector. The result shows that the RMB exchange rate due to the international capital flow and commodity trade two opposite the role and effect on CPI is not complete, the influence of stock market on CPI is not significant, the significant impact of foreign exchange on the CPI. Aimed at the impact of foreign exchange for the consumer price index using the non parameter method for more in-depth analysis. The results found in the influence of the different periods of CPI by foreign exchange is different, the two and the relationship is not completely positive correlation. Finally, relevant policy recommendations are given.
The innovation of this paper: first, the parametric regression model, nonparametric regression model, compare the semiparametric regression model, summarizes their advantages and disadvantages and applicability, convenient to others. Second, the impact of foreign exchange for the consumer price index using the method of nonparametric kernel regression is studied.

【学位授予单位】：天津财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.6;F832.51;F224

【参考文献】