The Scenario Simulation Analysis of the EU ETS Carbon Price Trend and the Enlightenment to China
Yi Lan, Li Zhaopeng, Yang Li and Liu Jie
摘要：2017 年中国即将建立全国性碳市场，碳价作为碳市场发展的风向标，其走势情况是碳市场政策制定的重要依据。本研究通过引入Stock600 指数，石油、煤炭、天然气价格等碳价的重要外部影响因素，建立BP 神经网络模型对EUA 碳价历史数据进行学习，从而模拟各因素对碳价走势的影响机理；通过分别控制经济发展和能源消耗两类变量，设立6 种典型情景以模拟碳价未来的可能走势，为全国性碳市场是否适宜纳入不同经济发展和能源消耗水平的区域提供理论依据。研究结果表明：①经济中、高速发展情景下，碳价与能源消耗同向变动，且碳价变动幅度及水平基本一致。高能源消耗下的高、中经济情景适宜同时纳入碳市场；②低经济发展情景下，无论能源消耗情景如何，碳市场均难以实现价格发现功能；③即使经济发展类似，高能耗与低能耗情景下的碳价依然存在较大价差。若同时纳入碳市场，低能耗情景将不得不承受由高能耗情景所主导的高碳价。
关键词 ： BP 神经网络模型, 情景模拟, 全国碳市场
Abstract: China’s national carbon trading market is due to be established in 2017 soon. As the best practice to follow,EU ETS, the world’s largest carbon market arouses strong academic interests. As an indicator of how the carbon marketdevelops, carbon price which is the reason for the fluctuation, trend analysis and forecasting is particularly important for research. This study aims to simulate the EUA prices by developing a BP Neural Network model to study the EUA historical prices together with external factors such as Euro Stock600, Brent oil, coal and natural gas prices. Fixing either a set of economic development variables or energy consumption variables to set up six typical scenarios to forecast the future EUA prices. The study attempts to provide theoretical basis for judging whether it’s appropriate for China to considering different economic development and energy consumption unanimously for national carbon trading market.The results show: When energy consumption is consistent, economically developed areas and economically secondary developed regions are suitable for the carbon market, and the carbon price of high energy consumption is more robust; Economic development in a downturn, the carbon market can not realize the price discovery function; Even if economic growth is similar, the prices of carbon in high-energy and low-energy scenarios still have a big difference and low-energy scenarios always have to endure high carbon price dominated by high-energy scenarios.
Keywords:BP Neural Network Model; Scenario Simulation; National ETS
DOI:10.19511/j.cnki.jee.2017.03.003