Abstract
This paper divides the energy market into energy futures market and new energy stock market. At the same time, the closing price of Shenzhen carbon emission rights is used to represent the carbon market price, the energy futures composite index of China Securities Exchange is used to represent the energy futures market price, and the stock price of new energy listed companies is used to represent the new energy stock market price. VAR model and MSVAR model are used to empirically study the relationship between the three variables and the nonlinear relationship between them. VAR model results show that there will be more complex relationship among carbon market price, energy company stock price and energy futures price. MSVAR model shows that the energy futures market, new energy stock market and carbon market present nonlinear and structural changes, and MSVAR model can better explain the nonlinear relationship among the three markets than traditional VAR model.
Highlights
In recent years, with the global warming and frequent extreme weather, energy conservation and emission reduction have become the consensus of people all over the world
The carbon market price is represented by the closing price of carbon emission rights in Shenzhen, the energy futures market price is represented by the comprehensive index of energy futures of CSI, and the stock price of listed companies of new energy is represented by the stock market price of new energy
The nonlinear relationship among the three markets is empirically studied by using vector auto regression (VAR) model and MSVAR model
Summary
The traditional vector auto regression (VAR) model is as follows: yφyφy , ⋯ , φ y ε (1) Among them, p Represents the number of valid lag periods, φj is an n times n order coefficient matrix, εi is a random disturbance term that obeys Gaussian white noise distribution. VAR model often ignores the structural changes and nonlinear characteristics of time series. Referring to the relevant literature, it is found that the vector autoregressive model (VAR) of Markov switching proposed by Hamilton can make up for the defects that VAR model cannot find the structural changes and nonlinear characteristics of time series. MSVAR model is introduced in this paper
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
