The Research On Time-varying And Heterogeneous Nonlinear Spillover Effects In International Crude Oil Market

As the basic energy for modern industry,the price volatility of crude oil has a significant impact on economic growth and financial m arket.Based on the non-linear characteristics of crude oil price changes,this paper constructs a time series analysis framework of cross-integration of multi-domain methods.From the perspective of time-varying,heterogeneity and non-linearity,this paper investigates the characteristics of crude oil price volatility and its volatility spillover to the condition mean of its own price.At the same time,it takes the basic fact that international crude oil price change has different impact mechanisms on different financial markets as the core.From the point of penetration,the spillover mechanism and transmission mechanism of information triggered by oil price volatilitys and risk shocks on stock market and exchange rate market are discussed.The main results and innovative contributions are reflected in the following aspects:Firstly,based on the time-varying spillover perspective,we construct a TVP-SVM model to study the time-varying feedback effect of oil price volatility,finding that the oil price volatility has a positive feedback effect,that is,the oil price volatility has a significant positive time-varying spillover effect on its price change.At the same time,based on Bayesian estimation method,14 GARCH and SV family time-varying volatility models are estimated.The marginal likelihood and DICs criteria are introduced to identify the optimal model for estimating international crude oil price and related energy product price volatility.The results show that the SV family model performs better than the GARCH family model in estimating crude oil price volatility,especially the SV-MA model.Secondly,based on the perspective of two-way heterogeneity spillover,we construct a cross-quantilogram model of two-way quantile decomposition to study whether the volatility of crude oil price has heterogeneity spillover effect on the returns in the BRICS stock market,and to test whether the former has direction prediction for the latter.The results show that the volatility of crude oil price has strong spillover and direction predictability to the stock returns of BRICS countries,and the degree of direction predictability depends on whether these countries are net importers or net exporters of crude oil.Thirdly,from the perspective of structural non-linear asymmetric spillover,we extend the SVAR model to study the structural spillover effect of different oil prices shocks(oil price changes driven by global economic factors(aggregate demand shocks),and the factors driven by international crude oil mar ket factors(non-OPEC supply shocks,OPEC supply shocks and specific demand shocks)on Chinese stock investor sentiment.The results show that stock investor sentiment does not respond significantly to aggregate demand shocks and non-OPEC supply shocks,while the negative response to crude oil specific demand shocks and the positive response to OPEC supply shocks are very significant,respectively.It proves that investor sentiment is an important way for oil prices to affect the stock market.Fourthly,based on the perspective of non-linear threshold spillover effect,we construct a STVAR model to study whether there are differences in spillover effect of crude oil price shocks on stock return under different credit market conditions.Based on the sample data of total stock market and industry stock returns in the United States,we find that when the U.S.economy is in a normal credit state,the positive(negative)impact of crude oil prices reduces(increases)stock returns.On the contrary,when the U.S.economy is in a tight credit situation,the positive(negative)shocks of crude oil prices will increase(decrease)stock returns.In addition,the spillover effect of oil prices on stock returns is mainly in the short term rather than in the long term.Fifthly,from the perspective of time-frequency heterogeneous asymmetric spillover,we construct a frequency-domain causal Granger test and a quantile Granger causality test model.Chapter 7 studies the heterogeneous spillover effect between crude oil price changes and real exchange rate changes in ten major economies(namely,Britain,Canada,Brazil,Russia,Mexico,Norway,India,Japan,South Africa and South Korea).The results show that there are time-frequency differences in spillover relationship between crude oil price changes and real exchange rate changes,that is,in the short and long term,the interaction spillover relationship between oil price and exchange rate is heterogeneous.In addition,changes in crude oil prices have spillover effects on real exchange rate movements in other countries except Norway,especially at the tail end of the distribution of exchange rate movements.This shows that when the real exchange rate is in extreme appreciation or depreciation,oil price changes will have a significant impact on the real exchange rate.However,when oil prices are extremely high,the real exchange rate changes in most net oil exporting countries will have spillover effects on oil price changes,while they will be ineffective when oil prices are extremely low.In contrast,the real exchange rate of most net oil importing countries can affect the price of crude oil regardless of the price of crude oil.