Font Size: a A A

Research On Hedging Ratios Based On M-Copula-GJR-VaR Model In Gold Market

Posted on:2014-03-16Degree:MasterType:Thesis
Country:ChinaCandidate:M QuFull Text:PDF
GTID:2269330425486845Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Determining the optimal hedging strategy has always been one of the mostimportant topics in the field of financial risk research, and the effect of hedging strategymainly depends on the accuracy of hedging ratios estimation. The VaR risk measuremethod has two advantages. On the one hand, it focuses on the negative yieldsfluctuation which accords with investors’real psychological feelings; on the other hand,in this method, investors’ investment demand can be met by setting an appropriateconfidence level. Therefore, this paper uses VaR to measure risk, and constructs aM-Copula-GJR-VaR dynamic hedging ratios estimation model which considers theasymmetry of the spot and futures markets, the cointegration relationship and nonlinearcorrelation between them with the basic of risk minimization principle. Then, the articletakes Chinese gold market with attributes of financial and commodity as empiricalobject to estimate the optimal hedging ratios.This paper firstly constructs GJR model with error correction term to estimate theconditional volatility of spot and futures returns and determines the specific marginaldistribution according to the fitting results. Furthermore, it uses cumulative distributionfunction sequence in different markets to estimate the M-Copula function constructedby appropriate single Copula functions, this paper obtains the joint distribution betweengold spot and futures returns based on the M-Copula-GJR model, and then calculatesthe optimal hedging ratio. Finally, it compares the hedging ratios and hedging effects ofM-Copula-GJR-VaR model with CCC-GARCH-VaR model, DCC-GARCH-VaR model,Clayton Copula-GJR-VaR model and Gumbel Copula-GJR-VaR model.The empirical results show that M-Copula-GJR-VaR model provides the besthedging effects when compared with CCC-GARCH-VaR model, DCC-GARCH-VaRmodel, Clayton Copula-GJR-VaR model and Gumbel Copula-GJR-VaR model. Theproposed model can be used to avoid the price risk of spot market with relatively fewerhedging cost; And after more than five years’ development, the hedging effects ofChinese gold futures market vary among0.672and0.704, which means the Chinesegold futures market is still immature.
Keywords/Search Tags:Gold futures, VaR, Hedging ratio, Nonlinear correlation, M-Copula-GJR-VaR model
PDF Full Text Request
Related items