| 653 | 18 | 36 |
| 下载次数 | 被引频次 | 阅读次数 |
针对不可观测异质性非时变假设导致的删失变量偏差及推断无效问题,本文构建了贝叶斯隐马尔科夫异质面板模型,刻画了截面个体间的动态时变不可观测异质性,对经济系统环境中可能存在的隐性变点进行了诊断,并设计了相应的马尔科夫链蒙特卡洛抽样算法估计模型参数,对中国各地区的金融发展与城乡收入差距关系进行实证分析,捕捉到金融发展与城乡收入差距间长期稳定关系的隐性变化,发现了区域个体不可观测异质性存在的动态时变特征。研究结果表明,各参数的迭代轨迹收敛且估计误差非常小,验证了贝叶斯隐马尔科夫异质面板模型的有效性。
Abstract：For time-invariant unobserved heterogeneity assumption lead to censored variable bias and invalid inference problem,this paper constructed Bayesian Heterogeneous Hidden Markov Panel Model to characterize the dynamics timevarying unobserved heterogeneity between sectional individuals and diagnose the hidden changes existing in economic system environment,designed the corresponding Markov Chain Monte Carlo sampling algorithm to estimate model parameters,and analyzed the relationships between financial development and urban-rural income gap. The empirical research captured hidden changes in the long term stable relationships between financial development and urban-rural income gap,found out that regional individual unobserved heterogeneity existed dynamic time-varying characteristics. Results showed that the parameters iterative trajectory converges and the estimated error was very small,it verified the effectiveness of the Bayesian Heterogeneous Hidden Markov Panel Model.
[1]Page E.S.Continuous inspection schemes[J].Biometrika,1954,41(1/2):100-115.
[2]Chow G.C.Tests of equality between sets of coefficients in two linear regressions[J].Econometrica:Journal of the Econometric Society,1960,28(3):591-605.
[3]Quandt R.E.Tests of the hypothesis that a linear regression system obeys two separate regimes[J].Journal of the American Statistical Association,1960,55(290):324-330.
[4]王维国,王霞.基于贝叶斯推断的上证指数突变点研究[J].中国管理科学,2009,17(3):8-16.
[5]Baltagi B.H.,Feng Q,Kao C.A Lagrange Multiplier Test for CrossSectional Dependence in a Fixed Effects Panel Data Model[J].Journal of Econometrics,2012,170(1):164-177.
[6]Cappe O.,Moulines E.,Ryden T.Inference in Hidden Markov Models[M].Springer,2005.
[7]Carlin B.P.,Chib S.Bayesian model choice via Markov chain Monte Carlo methods[J].Journal of the Royal Statistical Society,1995,57(3):473-484.
[8]Inclan C.,Tiao G.C.Use of cumulative sums of squares for retrospective detection of changes of variance[J].Journal of the American Statistical Association,1994,89(427):913-923.
[9]Bos T.,Hoentrakul P.Estimation of Mean and Variance Episodes in the Price Return of the Stock Exchange of Thailand[J].Financial Risk and Financial Management,2002(16):535-554.
[10]Broemeling L.D.Bayesian Inferences about a Changing Sequence of RandomVariables[J].Communications in Statistics-Theory and Methods,1974,3(3):243-255.
[11]廖远甦,朱平芳.均值和方差双重变点的贝叶斯侦测[J].统计研究,2011,28(11):91-97.
[12]Chib S.Estimation and comparison of multiple change-point models[J].Journal of econometrics,1998,86(2):221-241.
[13]Chib S.,Jeliazkov I.Marginal likelihood from the Metropolis–Hastings output[J].Journal of the American Statistical Association,2001,96(453):270-281.
[14]Burgess R.,Pande R.,Wong G.Banking for the poor:Evidence from India[J].Journal of the European Economic Association,2005,3(2-3):268-278.
[15]叶志强,陈习定,张顺明.金融发展能减少城乡收入差距吗?——来自中国的证据[J].金融研究,2011(2):42-56.
[16]Shaoping W.,Zhigang O.Y.The Threshold Effect of the UrbanRural Income Disparity on Real Economic Growth in China[J].Social Sciences in China,2008,3(4),39-53.
基本信息:
DOI：10.19343/j.cnki.11-1302/c.2014.07.014
中图分类号:O211.62
引用信息:
[1]朱慧明,黄仁存,游万海.贝叶斯隐马尔科夫异质面板模型及应用研究[J].统计研究,2014,31(07):97-104.DOI:10.19343/j.cnki.11-1302/c.2014.07.014.
基金信息:
国家自然科学基金委员会创新研究群体科学基金项目“金融创新与风险管理”(71271075);; 国家自然科学基金重点项目“战略导入的投资决策与风险管理”(71031004);国家自然科学基金面上项目“基于贝叶斯分位回归的面板数据建模、算法及应用研究”(71171075);; 湖南省研究生科研创新项目“空间面板分位数回归模型及其应用研究”(CX2013B127)项目资助
2014-07-15
2014-07-15