空间面板数据模型的研究及其应用

发布时间：2018-04-21 12:43

本文选题：空间面板数据模型 + 空间相关性　；参考：《广西大学》2014年硕士论文

【摘要】：由于地域差异和时间变化的存在,产生了空间面板数据,这类数据能同时反映时间和空间两个方向上的变化规律,比仅考虑地域差异的截面数据和只关注时间变化的时间序列数据适用性更广泛.空间面板数据具有空间相关性,是非独立的,这导致经典统计模型不再适用,而空间面板数据模型是解决这一问题的重要模型之一.空间面板数据模型中加入了可反映空间相关性的空问权重矩阵,能够解决综合时间序列和横截面数据的问题.因此,研究空间面板数据模型具有极大的现实意义.本文研究空间面板数据模型的参数估计,空间相关性检验和在区域经济中的应用. 首先,研究空间面板数据模型的空间相关性检验问题.分别用拉格朗日乘子检验法(LM检验法)和似然比检验法(]LR检验法)进行三种相关检验,用于判断数据是否具有空间相关性.通过复合假设检验判断是否存在空间相关性,再进行后两种边缘假设检验,区分存在的空间相关性是属于空间滞后还是空间误差. 其次,研究了具有更广泛应用的空间误差面板数据模型的参数估计问题.证明了空间误差面板数据模型参数的拟极大似然估计(QMLE)的渐近性质.还通过数值模拟验证上述理论,结果表明当样本容量足够大时,模型的参数估计值越接近其真值,且服从正态分布. 最后,将空间面板数据模型应用于我国区域经济增长问题的分析中,选取我国31个省市2001-2010年的经济增长数据进行实证分析.研究结果显示,区域经济增长速度不仅依赖于地区自身条件,还与相邻地区的经济增长息息相关,即存在空间相关性.期初人均GDP,人口自然增长率,政府消费占GDP比重和资本的积累程度对区域经济增长的影响较为显著.
[Abstract]:Due to the existence of regional differences and time changes, spatial panel data are produced, which can reflect the changes in both time and space at the same time. The applicability of cross-section data is more extensive than that of cross-section data which only takes into account regional differences and time series data which only focus on time variation. Spatial panel data is spatially dependent, which makes the classical statistical model no longer applicable. Spatial panel data model is one of the most important models to solve this problem. Spatial panel data model can solve the problem of synthesizing time series and cross-section data by adding spatial weight matrix which can reflect spatial correlation. Therefore, the study of spatial panel data model has great practical significance. In this paper, the parameter estimation, spatial correlation test and application in regional economy of spatial panel data model are studied. Firstly, the spatial correlation test of spatial panel data model is studied. Lagrange multiplier test (LM) and likelihood ratio test (LR) were used to determine the spatial correlation of the data. The spatial correlation is judged by the compound hypothesis test and the latter two edge hypothesis tests are carried out to distinguish whether the spatial correlation belongs to the spatial lag or the spatial error. Secondly, the parameter estimation problem of spatial error panel data model with wider application is studied. The asymptotic properties of quasi-maximum likelihood estimators for spatial error panel data model parameters are proved. The results show that when the sample size is large enough, the parameter estimation of the model is closer to its true value, and the model follows the normal distribution. Finally, the spatial panel data model is applied to the analysis of regional economic growth in China, and 31 provinces and cities in China are selected to analyze the economic growth data from 2001 to 2010. The results show that the speed of regional economic growth is not only dependent on the conditions of the region itself, but also closely related to the economic growth of neighboring regions, that is, there is a spatial correlation. At the beginning of the period, the per capita GDP, natural population growth rate, the proportion of government consumption to GDP and the degree of capital accumulation have a significant effect on regional economic growth.
【学位授予单位】：广西大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O212.1

【参考文献】