Study On Security Market Noise Estimation And Influence

As studies of behavior finance, continuous finance, nonlinear finance, securitymarket microstructure, risk pricing theories got deep, the hypothesis of perfect andcomplete market were unloosen and developed, the noise study is getting more andmore attention. Noise study mainly focus on two aspects, firstly it is whether noiseexists in market and if so, how to discern and estimate it, secondly how noise affectssecurity market, some think it beats market effectiveness, so it should be lower, themarket much better, however in the same time noise is an very important condition formarket normally running, because market would disappear gradually if there is nonoise. However due to complicated components the question of noise estimation isdifficult to resolve, then it hinder related studies. This paper start with the tow aspectsof noise questions, first of all, the paper explore measurements of noise from threedifferent angles, which embrace trade, return and level (volatility),then discuss theinfluence on security market from capital pricing and price behavior. In details, thepaper would do things and get corresponding conclusions as follows:1) By introducing the noise trader into the trade tree this paper develop a newtrade arriving model based on EKOP, then this paper deduce the single-periodlikelihood function and multi-period likelihood function of trade arriving based onPoisson arriving theory, in order to get the effective estimating model of noise tradeproportion of security market. The results of HS300sample stocks’ estimation ofnoise trade show that the noise trade proportion of Chinese security market is0.2432,a bit higher than information trade proportion. In the same time the results ofparameters estimation indicates that noise traders are more optimistic, the informationefficiency is low.2）By introducing the noise factor into O-U process this paper constructs thethree-factor model of price diffusion process which contains a mean-reverting noiseprocess. Then this paper employ the Kalman filter method and base on EM algorithmto accomplish purposes of unknown parameters estimation and measurement of noisereturn.then succeeds to measure the noise level at return scale. Then this papermeasure the week noise return level of SH000001from1999.01.04to2012.02.24, theresults show that its noise level is from-23.005to83.51%, right skew and highkurtosis, analysis show investors rationality and regulation affect noise level. 3) On the point of the nonlinear, deterministic and chaotic phenomena existing insecurity market, this paper introduce the thought of noise level estimating methodbased on phase space reconstruction theory from Signal science into security marketand build corresponding estimating model. Then this paper takes the HS300index assample, after testing the nonlinear, deterministic and chaotic characteristics of HS300index from20100104to20101214by using high frequency data, then this paperestimate their noise level, the results show that the noise level of HS300index is21.55-65.40%, right skew, and existing complicated relation between noise and assetprice trend and market information.4) Base on full reviewing the asset pricing theories and models, this paper drawlessons from MPT, CAPM, APT, LAPM, BAPM s ideas and conclusions, in thesame time simulating the method of Fama-French three factors model, then build anew CAPM (NAPM) model, and to explore the pricing situation of China stockmarket by new model. The results of empirical study on sample of HS300s stocksshows that noise has relatively strong positive pricing ability, and stock has morenoise then get more risk return for compensation.5) Finally, this paper build the dynamic panel VAR model that embrace the noiseand price behavior of refection of market quality based on introducing a new effectivemethod of noise level estimation. The empirical results of the sample study by HS300constitution stocks show that there are complicated dynamic relationships betweenthem, such as the positive correlation between noise and illiquidity and unsymmetricalinformation, the negative between it and volatility and volume, but they are unstableon the elementary stage.