Study Of Behavioral Characteristics And Price Forecasting Of The European Carbon Emissions Market

In view of the bad influence which is brought by the greenhouse effect to the society and nature, the international community promotes carbon reduction by carbon emissions trading actively. However, affected by economy, finance, politics, environment, climate and other factors, the price of carbon emission permits fluctuates wildly. So, it’s very necessary and urgent to study behavioral characteristics and forecasting methods of the carbon prices. The research object of this article is the European Union Emission Trading Scheme.Firstly, this paper introduces the theoretical basis of this article briefly that is the traditional efficient market theory, fractal theory, chaos theory and price forecasting theory. On the basis of these theories, we review the related literature.Then, this paper analyzes the behavioral characteristics of the EU carbon market. Through a basic statistical analysis carried out on the EU carbon market, we find that the EU carbon market doesn’t meet the three major hypothesis of the efficient market theory. So, the efficient market theory is not applicable in carbon market. Afterwards, this paper tests the behavioral characteristics of the EU carbon market within the framework of the fractal and chaos theory. To be specific, we test the fractal properties of the carbon market based on the two basic characteristics of fractal sequences. The empirical results show that all of the three phrases have the statistical self-similarity but only the first phase has the long-term memory. We test the chaos of the carbon market by analyzing the two basic features and the topological structure of the chaotic attractor. The carbon market is a chaos because of the fractional correlation dimension, the positive maximum Lyapunov index and the result of the CR test.Finally, this paper builds a model to predict the carbon price. Because the multifractal model is effective in depicting the complex price volatility, we use the wavelet leaders to test the multifractal characteristics of the EU carbon market. In this way, we find the diverse scaling characteristics in the carbon yield sequences. Hereby, a price forecasting model is proposed in this paper which is called as Db3-GA-RBF. Finally, by comparing with other price forecasting models, we prove the superiority and effectiveness of this model in the carbon price forecasting.