Research On The Conversion Rate Of Collateral In Derivatives Market

In the transactions of derivatives market margin system play a very important role.Margin requirements is not only a insurance against the opponent's credit risk,but also the base of the existence of leverage.In a mature derivatives market,the margin can be pledged by other securities,including treasury bonds,foreign currency,fund shares,stock-index constituent stocks,and so on.Due to the different risks contained in different securities,it is inevitable to evaluate its value and have a haircut when it be used as collateral.The ratio of the conversion price to the original price is called the conversion rate,which is the focus of research on the deposit system,as well as the core of this paper.The conversion rate has an effect on the activity of the derivatives market.A high conversion rate allows the assets to be pledged at a high value and lowers the requirements for the participants.Contrarily,the low conversion rate will increase the burden on the participants holding positions,but the risk will be covered fully.Different subjects have different willingness to take risks in different economic environments,which is reflected in the different rates of collateral.The study on rate of collateral of pledge is focused on its market risk.Based on the stock index funds represented by the Shanghai Composite Index ETF,this paper summarizes the data characteristics of the yield of the Shanghai Composite Index before analysis,and then provides the basis and conditions for the follow-up model's adoption and improvement.In this paper,the t-distribution is used to replace the normal distribution as the manifestation of the Leptokurtosis of the Shanghai Composite Index.In order to reflect the Volatility Clustering and Long-Memory,the GARCH model is used to predict the fluctuation from the yield of the Shanghai Composite Index,and then we choose the EGARCH model to reflect the Leverage effect existing in the process of fluctuation.Finally,we introduce the extreme value theory tailor the yield,and use the generalized Pareto distribution method to find the maximum risk value on the condition of given confidence.Through the comparison in empirical results of the independent model,we found the strengths and weaknesses of each model.Different characteristics and distinct working principle from different models provide the space for improvement.Under the support of the Second Extreme Value Theorem,a new model,the EGARCH-t-GP model,is born after combining with the EGARCH-t model and the GPD model.Experiment shows new model has better results than the single model in all aspects,and not only can passed various tests in different confidences but also provides higher accuracy on prediction and better dispersion on risk.It has been proved that the new model has inherited the complementary advantages from its parent model succeeding to get rid of genetic malformations,which is regarded as a preferable method to deal with similar problems.