Font Size: a A A

Application Of Gmdh Structure Mutation Search Model And Empirical Research

Posted on:2007-06-07Degree:MasterType:Thesis
Country:ChinaCandidate:L ZhouFull Text:PDF
GTID:2209360185955793Subject:Quantitative Economics
Abstract/Summary:PDF Full Text Request
Structural change has become a more and more important area of econometrics research, as well as financial invenstmet area.But the literature addressing the issue is relatively scarce. This dissertation attempts to make some new breakthrough on structure change modeling. What is more, for China's stock market has not only the character of new market, but also the particularity caused by institutional change, analyzing structural change is significant to comprehend and master the stock market fluctuation features.After studying a lot of literature on issues related to structural change, this dissertation introduces Group Method of Data Handling (GMDH), modeling GMDH auto-searching structural change, to avoid the possibility that the traditional endogenous methods maybe falsely decide the most outstanding break-point. Because of high intelligence of GMDH modeling, the location and the break kind of break-point can be determined automatically, then the optimal model can be obtained. Furthermore, it is applied to define the event date and the length of the post-observation period in event window to improve the validity of event study.Considering the factor that there possibly exist multiple structural changes in economic series, by referring to the modeling idea of Lumsdaine and Papell(1997), this dissertation has constructed GMDH model of multiple structural changes. The modeling method can solve not only the problem of the number and the location of breaks which the multiple structural changes theory requires, but also take the break kinds of every structural change into consideration, which is likely to be of particular relevance in practice. Furthermore, selecting the criterion value of F incremental statistic may control the significance of the breaks in order to accommodate to variant investigation intentions.By analyzing the empirical evidences on Shenzhen Composite Index, there exists strong policy effect in China's stock market, and it is not semi-strong form efficient to react macroeconomic policy. Researching the reduction of state-owned shares may provide references for the share splitting reform nowadays.
Keywords/Search Tags:Structural Change, GMDH, Auto-Searching, Policy Effect, Stage Fluctuation
PDF Full Text Request
Related items