关键词: 石油市场, 碳市场, 风险溢出效应, EVT-Copula-CoVaR

Abstract: In the context of global energy conservation and emission reduction, countries around the world have adopted the carbon market as an important way to save energy and reduce emissions. The carbon market treats carbon trading rights as a commodity that can be publicly traded in the market.By limiting the amount of carbon credits contracted to each country, carbon credits are traded between countries with insufficient and excess demand, or between companies, with the aim of controlling the total amount of carbon emissions of contracted countries worldwide. As the burning of oil is the main cause of CO₂ emission rise, there is an inextricable relationship between the price of oil and the price of carbon. The price of oil affects companies' demand for oil, which in turn affects carbon dioxide emissions, causing a change in the demand for carbon credits, which is ultimately reflected in a change in the price of carbon. This results in risky loss events in the oil market quickly spreading to the carbon market, meaning that the oil market can have a risk spillover effect on carbon market. Therefore, the analysis of the link between the oil market and carbon market should receive the attention of both theoretical and practical circles. In particular, financial market return data are characterized by spikes and thick tails. While traditional empirical distribution methods used to estimate marginal distributions tend to fit most observations, they poorly do so for the tail values, which are the risk extremes of greatest concern in risk measurement. At the same time, most current correlations between financial markets exhibit non-linear characteristics. Therefore, scientifically measuring the risk spillover effects of domestic and international oil markets on China's various carbon trading markets can not only enrich and improve the theoretical system related to risk spillover effects, but also provide policy guidance for the Chinese government to target the construction and development of carbon markets.
Based on this, this paper constructs an EVT-Copula-CoVaR model by combining the EVT, which fits the marginal distribution of the financial series, with the Copula function, which describes the dependency relationship, applying the CoVaR method to the study of the risk spillover effect of the oil market on the carbon market, and taking into account the extreme risk conditions of the market. This allows the risk spillover effect to be directly translated into a specific value. In particular, this paper quantifies the direction and intensity of risk spillovers from the oil market to the carbon market by selecting the price data of domestic and international oil markets and five actively traded Chinese carbon pilot markets from 2 April 2014 to 31 October 2019 as the research sample. We have drawn the following research conclusions. From the calculation of risk spillover effects: First, risk events in both domestic and international oil markets have a positive spillover effect on each pilot carbon market. This means that when a risk event occurs in the oil market, the risk in the carbon market increases accordingly. Secondly, a comparison of the different carbon pilot markets shows that the intensity of risk spillover from the oil market to the carbon market at the same level of confidence is in descending order: Hubei, Guangdong, Shenzhen, Beijing and Shanghai. The reason for this may be that the Hubei carbon market has the largest volume and turnover, and is therefore more susceptible to the influence of the oil market. Thirdly, comparing the spillover effect of domestic and foreign oil markets on the carbon market, it can be found that the spillover effect of foreign oil markets on the carbon market is greater than the spillover effect of domestic oil markets on the same carbon market. This may be due to China's increasing openness to the outside world in recent years, which has led to a yearly increase in the average daily oil import volume, surpassing that of the United States and thus becoming the world's top oil importer. As domestic oil dependency continues to rise, China's carbon market is relatively more influenced by the international oil market. In terms of the validity of the risk measurement model, the VaR forecasts for the domestic and international oil markets and the five carbon markets, as well as the CoVaR forecasts for the ten market combinations, are all within the critical range of the failure frequency test in terms of the number of days to failure. This means that the measures of VaR values for the seven markets and CoVaR values for the ten market portfolios are all valid, indicating the accuracy of the EVT-Copula-CoVaR model constructed in this paper and the Monte Carlo simulation method used for the risk measures.
Future research can build on this paper by introducing a time-varying Copula model to characterise the dynamic correlation between the oil market and the carbon market, so as to more accurately measure the risk spillover effects of the oil market on the carbon market.

Key words: oil market, carbon market, risk spillover effect, EVT-Copula-CoVaR model