Research On Irrational Factors Of IPO Premium Under Investment Bubbles

In June 2015,China's stock market reached a new record after one year prosperity,where Shanghai Stock Index reached a new record of 5178.19 points,the GEM index reached the level of 4037.96,the highest point in history.The market index increased several times in less than one year's time.With the market soared,many IPO also showing a high premium phenomenon.For example Storm technology,it staged a continuous daily limit myth since it land on GEM in March 24,2015.The opening price of storm of technology is only 9.43 yuan on the first day,in just two months,its stock price soared,up to 300 yuan,more than 30 times than its opening price.The paper raise the question in combination with the recent GEM market premium on the basis of the traditional theory of high investment bubble.Whether there is a high speculative bubble on our GEM.Which factors is relation with GEM IPO high premium.For different levels of foam,whether the differences between the factors of IPO underpricing.On this basis,the article using economics,econometrics,finance and other methods.Combined with GEM data,the paper conducted in-depth research with the relationship between the GEM IPO market bubble and high premium phenomena.Research ideas of this paper is divided into three levels.The first level focuses on the GEM market bubble of the first test.Analyzing the market bubble from earnings and Q ratio.According to results of the analysis,raising the hypothesis.There is the highest degree of bubble in GEM and serious speculative bubble from 2014 to 2015.Then applying exponential smoothing transfer model(GSADF)to test the existence of the GEM bubble.And draw the conclusions based on the results of the test.From 2009 to 2014,there is a lesser extent of bubble on the GEM market.In June 2014 to June 2015 GEM market bubble raise a higher degree.The second level is the factors of the IPO underpricing.Combined with the premium classical theory.The paper select four levels in investor sentiment,a positive feedback effect,inhibition of faith,the degree of market bubble,a total of eight feature variables as explanatory variables for initial gains.And make assumptions from the correlation relationship between characteristic variable and IPO initial return.Then conducting empirical analysis by stepwise regression method to the factors which affecting IPO high premium.Screening the characteristic variables which test results remarkable.Draw a conclusion that Investor sentiment,positive feedback effect,inhibition of faith,the degree of market bubble four factors were correlated with the degree of IPO Premium positive,fully illustrated that row irrational investors is a major cause of the initial IPO proceeds.The third level is discussed at different times,the main consideration of different degree of market bubble.The main difference between the factors of IPO underpricing.Construction of large data by using a second-level algorithm models derived characteristic variables-Random Forest model.Ranking variable factors in different periods(first-level derived)on the effect of IPO premium.The results concluded by the rank.Different times,different level of market bubble level and there is a big difference between the main factors of IPO Initial Return.GEM bubble level in 2009 to 2012 are relatively small,and the market is relatively more rational.The initial impact of IPO proceeds primarily from irrational factors in investors' optimistic sentiment and pre-market impact of positive feedback.In 2014 and 2015 GEM relatively high levels during the bubble,effect of IPO initial return of irrational factors mainly reflect IPO bubble about the degree of short-term market gains and different expected future stock between investor bullish or bearish price level.The structure of this paper is mainly composed of six parts,Specific arrangements are as follows:The first chapter is the introduction,which mainly introduces the background,research significance,framework,reference literature and possible innovations.The second chapter is the theoretical knowledge,including the detailed introduction of the stock market bubble theory and IPO irrational premium theory.In addition,This chapter defines and demonstrates the correlative concepts.The third chapter is the research hypothesis that there is the existence of the bubble of the gem,the correlation between premium rate and initial return,the differences of irrational factors in different periods,which is the foundation of model construction and empirical study.The fourth chapter is the model construction,including the GSADF model,the stepwise regression model and the random forest model,which provides support for the following empirical analysis.The fifth chapter is the empirical analysis,which is the key content of the article.According to the fourth chapter of the model and the third chapter of the hypothesis,this part bases on the gem data to make the empirical analysis and comes to the conclusion.Although there are a lot of discussions about the IPO premium of the GEM market,there are three possible innovations in this paper:First,because the traditional bubble inspection method is the disposable inspection process,There is not significant phenomenon of the test results of rupture bubble periodically.This paper uses the GSADF test proposed by Phillps to Recursive testing of the sequence cycle,which can overcome the problem of bubble rupture periodically.The model of the traditional IPO irrational premium only considers the correlation analysis between the influence factors and the premium rate,and the effect of the different influence factors on under different conditions has hardly been considered.This paper introduces the bubble factors on the irrational premium model of the traditional IPO to analyze the changes of the effect of IPO premium factors at different bubble levels,which provides a new perspective for the study of IPO irrational premium.This paper use machine learning model(the random forest)method to evaluate the effect of IPO premium variable and make tht Classification and regression analysis of premium variable through computer fitting with high accuracy in the sample which breaks the traditional statistical process(Assumed distribution-mathematical model fitting-hypothesis test-P test)and provides a new ideas for the research of IPO premium model.