Study On Markov Switching Model And Application In Business Cycle Analysis

The Markov-switching model proposed by Hamilton in 1989 is one of the most porpular nonlinear models at present.The model include serval structural equations so that it can be used to describe the time series' variate and transitions under disparate states.As the same reason analyzing the financal and macroeconomic variables with the Markov-switching model is a pop study area in the economics.There are many successful application in studying macroeconomic and financal variables with Markov-switching model, but as the difficulty of the model's estimate and some limitations in model's testing the model's popularizing is relativly not enough.Also the relative study on domestic economic problems with Markov switching model is very exility. The paper does some innovatie works on the method of model's testing and the applications in analyzing Chinese business cycle.The paper has six chapters. The background of topic, the mostly content, the method of study and the innovations of this paper are introduced in the first chapter of the paper. In the second chapter the author systematically describe the correlative theory of model's specification,estimating and forcasting etc.At first of the third chapter the model's specification testing which was proposed by Hamlton in 1996 is introduced.Also in the third chapter ,the Hansen's and Garcia's methods of testing for switching parameters were systematically expounded.The author firtly analyse the state variable's distrubiton character under the null hypothesis in the forth chapter,and according to the analysis the author propose aτstatistics.Then theauthor enduce theτstatistics' empirical distribution under different value ofβand proposed a new method of testing for switching parameters. At last the author also study the Hamilton's model for Ameican GNP with the new testing method,and get the result which indicate that the Hamilton's model is reasonable.In five chapter,for the first time, the author take a dummy variable into the traditional Markov-switching model to depict the change of Chinese business cycle pattern and Regime-Switching mechanism.The author analyze the Chinese business cycle using the modified Marlov-switching model.In the last chapter,the author summarize the paper and forcast the future works about the relative study.The innovations of the pape as follows: 1.For the first time in domestic, the author systematically describe and summarize the relative theorys about the Markov switching model; 2.A new method to generating the Markov-switching time series by Mont Carlo approach is proposed in the paper; 3.The author according to the characters of state varibale's distrubiton proposed a new method of testing for switching parameters,and analyzed the corrilation betweeenτstatistics' empiricaldistribution and the value ofβ; 4.At the first time,the author modify the thetraditional Markov-switching model according to the character of Chinses macroeconomic and analyze the Chinses business cycle using the modified Markov-switching model.