Oil Market Risks And Their Spillovers: Dynamic Copula-CoVaR Models

This paper mainly employs dynamic Copula-CoVaR models to research the oil market risks and their spillovers.In the methodology,this paper mainly solves the following four problems:? Referring to Joe(1997),Patton(2012)and Liu et al.(2017a),this paper constructs half-rotated Clayton and half-rotated Gumbel,and then makes up for the defects that the most common Copulas cannot describe asymmetric negative dependence and negative extreme tail dependence.Meanwhile,it derives the conditional Copula distributions and the Copula densities in details.? It derives the relationships between half-rotated Copulas and Kendall ?dependence,particularly defines two new tail dependences,i.e.lower-upper tail dependence and upper-lower tail dependence,and derives the relationships between the two new tail dependences and Copulas,and so makes up for the defects that the lower tail dependence and upper tail dependence cannot describe the extreme co-movement across financial markets or assets.Meanwhile,it summarizes the dependence properties for Copulas,and emphasizes employing half-rotated Copulas to measure asymmetric negative dependence,lower-upper tail dependence and upper-lower tail dependence.? It proposes a new driven variable to builds the dynamic process or regime-switching dynamic process for the dependence parameter of half-rotated Copulas,and then constructs a new dynamic copula model or regime-switching dynamic copula model to solve the problems that the most common dynamic copula model cannot capture the dynamic asymmetric negative dependence across financial markets.? It extends the definition of A-B CoVaR and G-E CoVaR to comprehensively describe the systemic risk or risk spillovers,and then focuses on how to measure the A-B CoVaR and G-E CoVaR.In particular,it emphasizes employing the conditional Copula distribution,i.e.Copula quantile,method to measure the A-B CoVaR and employing Copula joint distribution to capture various tail dependences to measure the G-E CoVaR.In the empirical research,this paper mainly conducts the following four works:? It researches the co-movement between oil market returns and uncertainty index changes under the extreme market risk perspective,and the empirical results show that there exists negative dependence between oil returns and uncertainty changes,meaning that the oil return will decrease when the uncertainty increases.Meanwhile,there are significant differences among the dependences for oil returns and various uncertainty indexes,and oil return risk responds asymmetrically to the extreme uncertainty indexes.? It compares the dynamic Co VaR of stock markets conditional on various oil shocks being extreme risks for countries closely related to oil price shocks,and the empirical results show that the dependence magnitude and risk spillovers highly depends on the stock market maturity as well as the type of oil shocks.Compared with the oil supply shocks and global aggregate demand shocks,the risk spillovers from oil market specific demand shocks to stock markets of BRICS and GCC countries are most obvious.? It analyses the Co VaR of currency markets conditional on oil market being extreme risks for oil trading countries from negative extreme dependence perspective,and the empirical results show that the dependences between oil returns and currency reutrns are weakly negative for oil trading countries including oil importing countries and oil exporting countries,i.e.the oil prices go up accompanied by the currencies of oil trading countries appreciating against US dollars.Meanwhile,the Co VaR further verifies that the oil market risks are determining factors of currency market risks,and the risk spillovers from oil market to currency market are asymmetric.? It employs dynamic model with regime switching to research the structural changes in the dependence between oil market and precious metal market,and the empirical results show that the low dependence regime exists between Brent crude oil market and gold market during the period from 2006 to 2012,and does not exist in other pairs.Meanwhile,the risk spillovers from oil markets to precious metal markets are significant,and the spillovers are larger when oil markets are in extreme downward risks.