| 656 | 1 | 71 |
| 下载次数 | 被引频次 | 阅读次数 |
传统固定效应变系数面板模型因在捕捉截面个体的异质性方面具有显著优势而被广泛应用于经济分析中,但其劣势在于忽略了不同截面个体之间的共同性。为弥补上述不足,本文构建了趋势固定效应函数系数面板模型,从理论上分析新模型相对于传统固定效应变系数面板模型的优越性,在此基础上针对该模型提出参数的两步估计法,并推导估计量的渐近性质。一系列蒙特卡罗模拟结果表明,本文构建的模型比传统模型具有明显优势,且提出的估计方法具有良好的有限样本性质。将模型和估计方法应用于外商直接投资(FDI)对地区经济增长影响的研究,结果表明FDI对经济增长的影响随着地区初始经济发展水平的变化而变化,即FDI对经济增长具有非线性动态偏效应。
Abstract:The traditional variable coefficient panel model with fixed effects is widely used in economic analysis because of its advantages in capturing the heterogeneity of cross-sectional units, but it ignores the commonality between different units. In order to make up the shortcomings, this paper builds a functional coefficient trending panel data model with fixed effects to theoretically analyze the advantages of the new model over the traditional variable coefficient panel model, and proposes a two-step estimation method. The asymptotic properties of the estimator are established. A series of Monte Carlo simulation studies show that the model constructed in this paper has the advantage over the traditional model, and the estimation method has good finite sample properties. Then we apply the model to study the impact of Foreign Direct Investment(FDI) on regional economic growth, showing that the impact of FDI on the economic growth changes with the different initial economic levels, indicating that FDI has a nonlinear dynamic partial effect on economic growth.
[1]白俊红,吕晓红. FDI质量与中国经济发展方式转变[J].金融研究, 2017(5):47–62.
[2]白仲林.面板数据计量经济学[M].北京:清华大学出版社, 2019.
[3]郭熙宝,罗知.外资特征对中国经济增长的影响[J].经济研究, 2009, 44(5):52–65.
[4]马岩.外商直接投资对我国经济增长的效应[J].统计研究, 2006(3):51–55.
[5]徐秋华,张梓玚.具有截面相关的变系数面板数据模型的估计及应用[J].系统工程理论与实践, 2019, 39(4):817–828.
[6]杨嬛.时间、节奏与FDI集聚的溢出效应[J].经济管理, 2019, 41(1):22–38.
[7]姚树洁,冯根福,韦开蕾.外商直接投资和经济增长的关系研究[J].经济研究, 2006(12):35–46.
[8]钟昌标,黄远浙,刘伟.外资进入速度、企业异质性和企业生产率[J].世界经济, 2015, 38(7):53–72.
[9] Bruce E H. Sample Splitting and Threshold Estimation[J]. Econometrica, 2000, 68(3):575–603.
[10] Cai Z, Chen L, Fang Y. A Semiparametric Quantile Panel Data Model with an Application to Estimating the Growth Effect of FDI[J]. Journal of Econometrics, 2018, 206(2):531–553.
[11] Cai Z, Fan J, Li R. Efficient Estimation and Inferences for Varying-coefficient Models[J]. Journal of the American Statistical Association, 2000,451(95):888–902.
[12] Chen J, Gao J, Li D. Semiparametric Trending Panel Data Models with Cross-sectional Dependence[J]. Journal of Econometrics, 2012, 1(171):71–85.
[13] Constantina K, Thanasis S. Foreign Direct Investment, Human Capital and Non-linearities in Economic Growth[J]. Journal of Macroeconomics,2010, 32(3):858–871.
[14] Fan J, Zhang W. Statistical Estimation in Varying-Coefficient Models[J]. Annals of Statistics, 1999, 27(5):1491–1518.
[15] Feng G, Gao J, Peng B, et al. A Varying-coefficient Panel Data Model with Fixed Effects:Theory and an Application to US Commercial Banks[J]. Journal of Econometrics, 2017, 196(1):68–82.
[16] Gao J, Xia K, Zhu H. Heterogeneous Panel Data Models with Cross-sectional Dependence[J]. Journal of Econometrics, 2020, 219(2):329–353.
[17] Juan M R, Alexandra S. Efficient Nonparametric Three-stage Estimation of Fixed Effects Varying Coefficient Panel Data Models[J]. Statistica Sinica, 2021(31):981–1003.
[18] Juan M R, Alexandra S. Nonparametric Estimation of Fixed Effects Panel Data Varying Coefficient Models[J]. Journal of Multivariate Analysis,2015(133):95–122.
[19] Juan M R, Alexandra S. Direct Semi-parametric Estimation of Fixed Effects Panel Data Varying Coefficient Models[J]. The Econometrics Journal, 2014, 1(17):107–138.
[20] Li D, Chen J, Gao J. Non-parametric Time-varying Coefficient Panel Data Models with Fixed Effects[J]. The Econometrics Journal, 2011, 3(14):387–408.
[21] Pei Y, Huang T, Peng H, et al. Network-based Clustering for Varying Coefficient Panel Data Models[J]. Journal of Business&Economic Statistics, 2022, 40(2):578–594.
[22] Pei Y, Huang T, You J. Nonparametric Fixed Effects Model for Panel Data with Locally Stationary Regressors[J]. Journal of Econometrics,2018, 2(202):286–305.
[23] Richard B, Henrik B, Rikard F. Multinationals, Endogenous Growth and Technological Spillovers:Theory and Evidence[J]. Review of International Economics, 2005, 13(5):945–963.
[24] Sun Y, Raymond C, Li D. Semiparametric Estimation of Fixed Effect Panel Data Varying Coefficient Models[J]. Advances in Econometrics,2009(25):101–129.
[25] Ullah A, Roy N. Nonparametric and Semiparametric Econometrics of Panel Data[M]. In Ullah A, and Giles D E A(Eds), Handbooks of Applied Economic Statistics. New York:Marcel Dekker, 1998:579–604.
(1)实际上,本文使用的函数系数模型是传统变系数函数的一种拓展,为了将本文提出的模型与传统变系数面板模型区别开,称模型中偏效应是函数形式的面板模型为函数系数面板模型。
(1)原因是在使用局部多项式法估计模型时,差分或组内变换的处理方式会使得变化后的模型中非参数成分增加,会进一步使得核函数权重的维度过高从而导致估计量的方差变大。
(2)由Sun等(2009)首次提出了在估计固定效应变系数面板模型时,结合传统线性面板模型中的虚拟变量法转换了固定效应之和为零的可识别性条件。
(3)在后文结合虚拟变量法的估计步骤中消除了该可识别性条件。
(4)为了估计趋势项有必要将固定效应估计出来,在实际应用中,针对该模型一般只关注函数系数和趋势项的估计结果。
(1)因篇幅所限,估计过程以附录1展示,见《统计研究》网站所列附件。下同。
(2)因篇幅所限,第2步估计的具体过程以附录2展示。
(1)因篇幅所限,定理2的证明过程以附录3展示。
(2)当Zit为多维向量时,模拟情况类似。
(3)因篇幅所限,模拟1~7的结果以附表1~9展示。
(4)当经济变量之间的变化存在周期性波动时,三角函数能够捕捉其周期成分;当某指标受到冲击后有短时间的爆发性变化时,指数函数能够反映其巨大的变化幅度;对数函数能够描述先快速增长后缓慢增长的变化形式;当变化趋势随某一变量的变化而变化时适用于多项式函数。
(1)因篇幅所限,■的AMSE结果以附表10展示,■与真实值的对比图以附图1展示。
(2)结合模拟1,解释变量生成过程为, ■,并将结果进行对比分析。
(3)模拟1中数据生成过程使用了Epanechnikov核函数,可结合模拟1中函数系数为m(Z)=sin(0.25πZ)的估计结果分析。
(1)固定资本形成总额在2017年之后不再公布,因此2017—2021年的数据根据固定资产投资的增长率近似计算得到。
基本信息:
DOI:10.19343/j.cnki.11-1302/c.2024.08.011
中图分类号:O212;F832.6;F124
引用信息:
[1]刘汉中,刘可.趋势固定效应函数系数面板模型的估计与应用研究[J].统计研究,2024,41(08):150-160.DOI:10.19343/j.cnki.11-1302/c.2024.08.011.
基金信息:
国家社会科学基金重点项目“内生性资源错配的形成机理及其对全要素生产率的影响研究”(18AJL004); 广州大学研究生创新能力培养资助计划“变系数面板模型理论方法及应用研究”(2021GDJC-D01)
2024-07-25
2024-07-25
2024-07-25