Theoretical Research Of Discrete Selection Model And Its Applications

As a usually used method in microeconomics and macroeconomics, in the paper we firstly listed systematically the theoretical methods of different types of discrete choice models, and then did some research works on the theoretical problems,at last, we gave some applications of discrete choice model.The contents of the paper can be divided into seven chapters.Chapter 1,which is the Foreward, we introduced the writing background of the paper, the reviews of the discrete choice model,the research problems and the innovations of the paper;Chapter 2 makes a reviews of the discrete model and Bayes Economitrics. In Chapter3,after listing systematically the estimation methods of the discrete choice models using the crosssection data, we analysised the impacts of different algorithms and prior specifications to the properties of the estimator.At the same time,we compared the effectiveness of the estimators and the accurateness of hypothesis tests of different methods belong to Frequency School and Bayes School;In Chapter4,we reviewed the theoretical methods of discrete choice models using time series data, proposed some problems,and then analysised the convergence and its rate of the MLE when the explanatory variable is a random walk process with a drift in the binary choice model; In Chapter5,we generalized the commonly used methods for the fixed and random effects discrete choice models of panel data;In Chapter6,using the methods of Bayes econometrics and the discrete choice models,we analyzed some problems in the reality;In Chapter7,we summarized the paper and brought forward some problems that can be considered in the future.We focused on the theoretical analysis in the paper.we mainly did three works in the paper:Firstly, given the binary probit model,we analysed the effects of different algorithms and prior specification on the properties on the estimator.The results revealed that:When the sample size is small,"biased" prior specification can lead the Gibbs sampler estimator be biased, but the phenomenons can't be happened for the M-H and Rejection sampler estomators.When the sample size is large,the "bias"of the prior specification has no affect on the unbias of the estimators; Secondly, given the MNL and ordered Probit model,when the prior distributions of the parameters are flat, we found that from the estimation and hypothesis test point of view in small sample size, the methods of Bayes School are more priorior than that of Frequency School; Thirdly, in order to discribe the obvious trends in macroeconomic time series, we put the random walk procee with a drift into the binary model,using the "response surface function", we analysised the the convergence and its rate of the MLE of the model, the Monto Carlo similations revealed that the MLEs are consistly and unbaised, the rate of convergence is T4/3.Empirical analysis included the following four aspects:Firstly, In the context of the population of Pandemic Influenza H1N1 in the world, we think of the possible mixed effects, independent effects, correlated effects and spatial correlated effects of the mortality rates of Pandemic Influenza H1N1 in different countries. Using the model choice methods in Bayesian econometrics, we decide the most appropriate model to discribe the mortality rates of different countris. Base on the model, we estimate the mortality rates of different countris. The result reveals that:the individual independent and spatial correlated effects model can best fit the sample data, and the estimated average mortality rate of Pandemic Influenza H1N1 in the world is 0.577%; Secondly, in order to ananlysis the impact of households'behavior selections on rural poverty in China, after defining the degree of rural poverty reasonably, using the suevey data, we employ an ordered probit model to find whether the households' behavior selections can influence the degree of rural poverty. the results reveal that: the numbers of enterprise, the efficiency of rural management, the employment structure, the degree of education of the households, the public investement and the state of popularization of agricultural technology all determinate the degree of rural poverty in China; ThirdlyIn the context of increasing volume in the inter-bank loan market, the offered real interest rate fluctuated violently, so the commercial banks facing the interest rate risk.In order to reduce the risks they faced, we analysised the determinants of the adjustments of the offered real interest rates in the inter-bank loan market.The results revealed that:all the changes of the government's expenditure, the investment of the fixed assets and the money supply can influence the adjustments of the real interest rate significantly. however, both the changes of people's consumer expenditure and the import and export commercial trading can't.The findings reflect the effectiveness of our macroeconomic policy in the currency market.