On The Market Microstructure Of China’s Gold Market

The study about China’s gold markets in this paper derives from two aspects. One aspect is, in a broad sense, to theoretically make a detailed, comparative analysis between Tianjin Precious Metals Exchange (TJPME thereafter) and Shanghai Gold Exchange (abbreviated as SGE) from the perspective of market microstructure. The analysis brings to a conclusion that the imperfection of the market microstructure inTJPME results into a variety of risks. The other is to employ the game theory to analyze the market access mechanism of gold markets from the prevention of credit risk. A conclusion comes that the membership should exist in both exchanges, and that the differences between TJPME and SGE make sense.Particularly, the imperfection of the market microstructure inTJPME involves three regimes, namely, market access mechanism, pricing mechanism, and clearing mechanism as well. The imperfection of market access mechanism is characteristic of the qualifications of general members, market makers and special members; The imperfection of pricing mechanism characteristic of uncertain stocks, shadow price, private information and leverage difference; The imperfection of clearing mechanism characteristic of delivery settlement and clearing risk. Hence, TJPME is characterized by such risks as credit risk, liquidilty risk, asymmetric information, and market risk. In this paper, the non-repetitive game is applied to analyze the loss when the participants discard some opportunities for the sake of credit risk. Consequently, the one-sided collateral makes no sense, whereas the double-sided collateral should be taken into account. The collateral of one party maybe returned either to himself, or to the other party, which represents the vital importance of the qualifications for all the members. The rationality and necessity of SGE’s credit risk control results from its members’ variety and ranking. That of TJPME’s membership is a question, however.Some market microstructure theories and econometrics concerning time series are employed to make an empirical analysis of market performences. The empirical implications are listed as follows:Firstly, three cases are applied to study the liquidity performance of the AU(T+D) species in SGE. These cases are to put into market gold futures in Shanghai Future Exchange, to put into market TJPME’s gold species, and to cancel it as well. The first case reveals that the liquidity performance of the AU(T+D) species can be promoted in the long run. The second one shows that the appearance of TJPME’s gold species exercises a complicated impact on that of the AU(T+D) species, while a positive impact exists in the long run. The third reveals that the cancellation of TJPME’s gold species also has a similar influence on its liquidity performance.Secondly, some low frequency data are employed to make an empirical study about the volatility performance of the AU(T+D) species. The empirical implications are:(1) such GARCH patterns as GARCH(1,1), TGARCH(1,1), EGARCH (1,1) and EGARCH(1,1)-M are used to effectively explain the volatility performance of the AU(T+D) species, in which EGARCH(1,1)-M ranks No.1;(2) the asymmetric effect appears in SGE, an effect of the positively external impact(good news) stronger than the negatively external impact(bad news);(3) the price follows a positive principle of risk-return.Thirdly, some high frequency data are also used to make an empirical study about the volatility performance of TJPME’s gold species. The empirical findings are:(1) the returns under different time series show that they exhibit kurtosis and fat tails;(2)the intra-day time series reveal that the volatility of TJPME’s gold species is shaped as M;(3) the volatility of TJPME’s gold species displays the typical leverage effect with the negatively external impact(bad news) twice stronger than the positively external impact(good news), and TGARCH(1,1) is the best pattern for demonstrating its volatility performance;(4)the difference between realized volatility model and weighted realized volatility model makes little sense, and the optimal time span is10minutes.All these findings are extremely important, and make the writer figure out some proposals for promoting the liquidity and volatility performance of double gold exchanges.