A Study On Open-end Fund Risks Early Warning Model With Countercyclical Supervision

With the rapid economy growth, the security investment fund is developing wellin recent years and further pushes the economy. And then the risks come to light.Although we generally regard it is of theoretical and practical significance to identify,estimate, control and establish an early warning system (EWS) for security investmentfund, the present researches focus on risks identification and estimation with littleattention on an early warning system. Meanwhile, since the international financialcrisis in2008, more and more specialists and officers are aware of financial system’spro-cyclicality and instability, and call on the change from micro-prudential policiesto the macro ones, which may be much more useful for dispersing risk. However,most researches have noticed the pro-cyclicality in banking, not in security investmentfund. Thus we try to study the risks of security investment fund from the view of pro-and counter-cyclicality, and establish an early warning system.This dissertation is scheduled as follows. The first chapter is an introduction,including the background, significance, roadmap, and innovation point. The secondchapter reviews the literature, as well as the inadequacy of the current researches onopen-end funds risk and some ideas we could borrow form counter-cyclicalsupervision theory and financial risk EWS. Then the third chapter outlines thefeatures of open-end funds risk, and calculate CSI Fund Index VaR to establish aregression model with GDP growth rate and Shanghai composite index growth raterespectively. The fourth chapter puts forward an indicators system for open-end fundrisk’s EWS, which has been built on either Chi-square Automatic Interact Detector(CHAID) or Bayesian Network. The final chapter is conclusions, expectations andsuggestions.On one hand, the empirical results show us the pro-cyclicality between open-endfunds risk and GDP growth rate and Shanghai composite index growth raterespectively, which emphasize the necessity to build a counter-cyclical supervisionsystem for open-end funds. On the other hand, the EWS we have established owns aprecision of predicting of75%~85%, which means we can basically classify open-endfunds correctly; and the EWS base on Bayesian Network even is able to figure outeach node’s conditional probability. The innovation points of this paper are:(1) thinking of the open-end funds riskmanagement from the viewpoint of counter-cyclicality;(2) demonstrating thepro-cyclicality of open-end funds risk, which exemplify financial system’spro-cyclicality and instability;(3) building EWS based on CHAID decision tree andBayesian Network respectively, which is a great progress comparing with usualmethods.