Research On The Bubble In Stock Market

Based on some qualitative analyses of bubbles existing in stock markets, the dissertation investigates the complex dynamic process from appearance, evolution to the death of bubbles using nonlinear dynamic system and econophysics. At present, the main research method of bubbles in stock market is of a qualitative way rather than of a quantitative one. The research logic clue of this dissertation is from phenomenon window to explanation window. Firstly, the bubbles existing in an economic system, especially in the stock market, are described comprehensively. Then, empirical analyses of bubbles in Shenzhen and Shanghai stock market are made. Based on research results of economics and psychology, the psychological factors that affect the performance of a stock market price are discussed. From the perspective of a nonlinear dynamical system and econophysics, the dynamical evolution process of a stock market is explained. The critical point of bubbles crash is analyzed with a complex system model. At last, through Grand Canonical Minority Game (GCMG), an investigation into crash avoidance in a complex system is advanced and the method to avoid bubbles crash in theory is proposed. The main contents in this dissertation are as follows:1 The various ways of testing bubbles in stock markets are proposed. Using the data samples from 18 Dec. 1990 to 7 Jan. 2002(SZSI) and 3 Apr. 1991 to 7 Jan. 2002(SHCI) in China stock markets and the ways of dynamic auto-regression and excess volatility, an empirical evidence is provided, showing that the bubbles are ubiquity in despite of a current decreasing tendency in market bubbles. 2 The empirical analyses of relations between market bubbles and market efficiency are made. It has been found that the time when market bubbles appear is also the time the market efficiency disappears.3 From the view of economics and psychology, the psychological factors of investor that induce bubbles appearance in stock market are discussed.4 Based on imitation and contrarian behavior model and heterogeneous agent model, the dynamical process of the market price is described. It is found that the price behavior shows various patterns of speculative dynamics including speculative bubbles, speculative chaos, and large decaying swings away from the fundamental price under a certain condition.By setting up econophysics model, the characteristic of price variability before a5 crash is presented. Our analyses show that local imitation among noise trades might cascade through the scales into a global coordination and cause crashes.6 Using Grand Canonical Minority Game (GCMG) model, a possible method to avoid bubble crashes is proposed. It is found that a system can be "immunized" against large change: by inducing deliberately small changes today, much larger changes in the future can be prevented.