The Study On Theory And Application Of The Survival Model

The empirical analysis of Micro-Economic theory became more and more convenient and possible because of the establishment and development of large Micro-databases in recent years. the field Micro-Econometrics has been widely applied and rapidly expanding. In this paper, we focused on an important branch of Micro-Econometrics-the survival analysis.This paper is divided into three parts. The first part is the general introduction, namely chapter 1. In this part we introduced the selection of themes, literature review and structural arrangements of this article. The second part is the main body of this dissertation, mainly focusing on the theoretical aspects of survival analysis, including chapter 2 to chapter 6. Chapter 2 described the nature of the survival model and basic principles; chapter 3 and chapter 4 introduced the non-parametric, semi-parametric and parametric estimation methods and statistical inference; chapter 5 focused on the martingale-based counting process and the survival statistics of the asymptotic distribution; chapter 6 described the further expansion of the survival model. Chapter 7 is the third part of this dissertation, involved two applications, unemployment in the labor force and price continuity of the securities markets.Innovation of this dissertation includes the following four aspects:(1) The dissertation conducts a Monte Carlo experients on the coefficient distribution of parametric model for finite sample and gives the percentile of these distributions under different sample size. (2) The dissertation conducts a Monte Carlo experients on exponential distribution, Weibull distribution, lognormal distribution and log-logistic distribution to give the sample size beyond which the coefficient distribution reaches the normal distribution. (3) Two empirical analysis are conducted. First, applying the survival model on the unemployment problem labor force of 3500 households survey data in Shanghai, Shenyang, Wuhan, Xi'an and Fuzhou, this dissertation uses non-parameter, semi-parametric and parameter model, finding that over time, the probability of the unemployed labor re-get a job gradually decreases.If the unemployment duration is more than 120 months, it is hard to find a job. Whether participating in training process and marital status of unemployed workers significantly affects the re-employment probability of labers. The higher Level of economically developed status is favorable for high re-employment rate. Second, the nonparametric, semi-parameter and parameter model is applied to the persistence of stock price of 1425 companies on the Shanghai and Shenzhen Security Exchange after the disclosure of annual report in 2008. The conclusion shows that the survival function is different between price rising and declining after the disclosure of annual report. Generally stock price continued rising for a long time and falling for a short time. For pricing rising company, the return on sales, return on assets, revenue growth rate and operating cash flow growth rate are significant and the dummy variable for mining, manufacturing, gas, electricity and water production and supply, real estate and comprehensive industry are significant, which indicating the investors in China are more concerned with the financial information of profit and growth rate. For pricing declining company, the inventory turnover and operating cash flow growth rate are significant and all dummy variables are insignificant. Growth rate of operating cash flow are significant in both groups, which indicating investors are concerned about growth rate of operating cash flow.(4) The survival analysis has only been non-systematicly studied for a short time, mainly for simply applications. This dissertation comprehensively analyzes and summarizes the latest research progress based on systematically collecting the literature of survival analysis.