项目名称: 经济时间序列的状态转换研究及其应用
项目编号: No.71301072
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 管理科学
项目作者: 蒋彧
作者单位: 南京大学
项目金额: 23万元
中文摘要: 经济变量在其运行过程中受到突发事件的影响,运行态势会产生明显的改变,并且会在数种不同的运行状态之间进行更迭,从而导致经济时间序列具有状态转换的动态特征。近二十年来,时间序列的状态转换研究是计量经济学领域的一个热点问题,其中,运用马尔可夫链描述状态转换过程的马尔可夫转换模型在经济与金融领域得到了广泛的应用。本项目拟基于贝叶斯方法,构造马尔可夫转换模型研究经济时间序列的状态转换特征。首先,将着重研究模型比较和选择问题,包括最优模型的选择以及最优状态数目的确定标准;其次,项目将针对后验模拟中的标签互换问题,设计时间序列状态识别的方法。以上两方面研究将在一定程度上解决马尔可夫转换模型的现有问题。另一方面,项目将针对我国经济时间序列开展实证研究,确定经济变量运行的最优状态数目,明晰其运行的动态特征,掌握状态转换的内在演化机理,为我国经济与金融政策的制定提供重要的参考依据。
中文关键词: 马尔可夫转换模型;贝叶斯;状态数目;标签转换;经济时间序列
英文摘要: Affected by unexpected events, evolution of economic variables may experience significant changes. Furthermore, the evolution may exhibit transitions between several different states and therefore the corresponding economic time series show dynamic characteristics of state or regime switching. In the last two decades, regime switching of economic time series has become one of the important topics in econometrics. Among various models, the Markov-switching model that applies the Markov chain to describe the dynamics of regime switching has been widely used in economics and finance. The project will construct the Markov-switching models to study the regime switching of economic time series based Bayesian framework. First, we will focus on the issues of model comparisons and selections, including the selection of the optimal model and the criterion of the optimal number of regimes. Second, we will develop methods of state identifications regarding to the label switching problem in the posterior simulations. The study of these two aspects will to some extent solve the existing problems of the Markov-switching models. On the other hand, the project will carry out empirical studies of economic time series in China. We will first determine the optimal number of states in the evolution of various economic variables and
英文关键词: Markov-switching model;Bayesian;the number of states;label switching;economic time series