火电企业价格风险预测模型与对冲策略研究

发布时间：2018-05-06 22:24

本文选题：火电企业价格风险 + Copula-MSM-GARCH模型　；参考：《中国矿业大学(北京)》2017年博士论文

【摘要】：当前我国正处在电力市场化改革攻坚期和二氧化碳排放配额交易计划正式启动期。火电企业价格风险源从单一的煤炭价格波动转变为煤炭、电力和二氧化碳排放配额三种价格波动。三者波动规律都有着各自的特点,且互相影响、互相依赖。在此背景下,未来火电企业价格风险管理问题必将变得更加复杂。因此,煤炭、电力和二氧化碳排放配额三者各自的波动规律、彼此的相关性结构、集成风险预测、风险对冲等问题都亟待研究。本文为煤炭、电力和二氧化碳排放配额各自的波动规律建立了多种模型,包括马尔科夫转换多重分形(MSM)模型和经典GARCH模型,用MSE、MAE、SPA等波动率预测评价方法比较了各种模型设计规格的波动率预测能力。在众多的Copula函数中选择了学生t-Copula函数作为连接函数,新建了一个Copula-MSM-GARCH模型,并评价了该模型的波动率预测能力和最小CVaR投资组合构建能力。此外,本文还提出了以期货交易把现实生产中的商品组合调整为条件在险值(CVaR)最小投资组合的风险对冲策略。本文各章安排如下:前言部分首先限定了本文研究对象为火电企业价格风险,随后在研究背景与意义中介绍了中国正面临推进电力市场化改革和开启二氧化碳排放配额交易的时代背景,并说明了当前电力市场化改革和二氧化碳排放配额交易启动使火电企业面对的风险因素发生变化,急需建立更好的集成风险预测模型。在国内外研究现状中,介绍了国内外关于电力、二氧化碳排放配额和煤炭三种商品价格波动规律建模的相关研究,还介绍了国内外投资组合集成风险预测的相关研究。此外,引言部分还详细阐述了本研究的研究内容和所用方法,列出了本文的研究思路,以为撰写全文做准备。第二章研究电力价格波动规律。之前学者研究发现,电价具有多重分形性。本文通过对电价的多重分型模形的研究接受了该发现。本章针对电力价格的波动规律,建立了一个马尔科夫转换多重分形(MSM)模型。选取EEX数据进行实证检验。有关电力现货价格波动率预测表现的研究显示,多重分形模型有能力比GARCH模型在该方面做得更好。第三章研究二氧化碳排放配额价格波动规律。本章研究了EU ETS的二氧化碳排放配额的短期现货价格波动规律。在回顾了该类新兴资产的典型事实后本章研究了多种方法以建模排放配额波动规律。纵观不同阶段价格和回报率的行为,本章建议用马尔科夫转换模型和AR-GARCH模型进行随机建模。为了检验这些方法,本章进行了样本内和样本外预测分析,并比较了两种方法的预测结果准确度。结果是,该模型足以捕捉到如偏性,超额峰度和不同阶段波动率行为在内的主要特征。第四章研究煤炭价格波动规律。本章采用马尔科夫转换多重分形模型和一连串广义自回归条件异方差(GARCH)类模型来建模并预测煤炭价格波动率。本章延伸了魏宇等(2010)和王玉东等(2016)之前的研究[57-58],用预测能力优越性检验(SPA)评价了所有这些模型的预测表现。为了准确预测煤炭价格波动率,本章尝试应用了多种不同类型的MSM模型。本文还在考虑了波动率与VaR两种风险度量方法。通过比较其他研究所用模型和MSM模型的预测表现,本章确定,新的MSM模型在各种预测时域上都优于其他模型。该优越性也适用于VaR的预测。第五章研究火电企业集合价格风险对冲策略。利用条件Copula函数体系,本章可以分别建模相关性结构与各商品价格波动率。之前各章已经筛选出了煤炭、电力和二氧化碳排放配额价格的波动规律最合适的模型设计规格。在此基础上,本章找到了最适合连接煤炭、电力MSM模型与二氧化碳排放配额GARCH模型的学生t-Copula函数,从而建立了Copula-MSM-GARCH模型,并用欧洲能源交易所EEX数据做了实证。本章也展示了Copula-MSM-GARCH模型配合Monte Carlo模拟法如何被用于发现最小CVaR商品组合,并提出最小CVaR商品组合可以作为风险对冲策略的标杆,以期货交易把现实生产中的价格风险组合调整为最小CVaR投资组合可以使预期价格风险最小化。第六章为结论与展望。本章总结了全文的研究成果,提出了论文研究的创新点和研究存在的不足之处,并对下一步工作进行展望。
[Abstract]:At present, China is in the hard period of the reform of the electricity market and the official start of the carbon dioxide emission quota trading plan. The price risk sources of the thermal power enterprises change from the single coal price fluctuation to three kinds of price fluctuations of coal, electricity and carbon dioxide emissions. The three fluctuation laws have their own characteristics, and they interact with each other and depend on each other. In this context, the price risk management of thermal power plants will become more complicated in the future. Therefore, the fluctuation laws of coal, electricity and carbon dioxide emissions quotas, the correlation structure of each other, the integrated risk prediction, and the risk hedging are urgently needed to be studied. This article is for the coal, electricity and carbon dioxide emission quotas each of the three. A variety of models are established, including the Markoff transform multifractal (MSM) model and the classical GARCH model. The volatility prediction ability of various model design specifications is compared with the volatility prediction evaluation methods such as MSE, MAE and SPA. The students' t-Copula function is selected as the connection function in many Copula functions, and a new Co is built. Pula-MSM-GARCH model, and evaluation of the volatility prediction ability of the model and the minimum CVaR portfolio construction capability. In addition, this paper also puts forward the risk hedging strategy of adjusting the commodity portfolio in real production to the risk value (CVaR) minimum portfolio. The chapters are arranged as follows: the preface is first limited The object of this paper is the price risk of thermal power enterprises. Then, the background and significance of the research are introduced in the background and significance of China. China is facing the background of promoting the reform of the electricity market and opening the carbon dioxide emission quota transaction, and explains the risk factors of the current electricity market reform and the start of the carbon dioxide emission quota transaction. There is an urgent need to establish a better integrated risk prediction model. In the domestic and foreign research status, this paper introduces the relevant research on the modeling of three commodity price fluctuation laws of electricity, carbon dioxide emission quota and coal, and also introduces the related research on the integrated risk prediction of domestic and foreign investment portfolio integration. In addition, the introduction is also detailed. The research contents and methods used in this study are expounded, the research ideas of this paper are listed, and the full text is prepared. The second chapter studies the law of electricity price fluctuation. The previous scholars found that the electricity price is multifractal. A Markov switching multifractal (MSM) model is established. EEX data is selected for empirical test. Research on the prediction performance of electric spot price volatility shows that the multifractal model has the ability to do better than the GARCH model. The third chapter studies the price fluctuation law of carbon dioxide emission quota. After reviewing the typical facts of the EU ETS carbon dioxide emissions quotas, this chapter studies a variety of methods to model the fluctuation of emission quotas in this chapter. A survey of the behavior of different stages of price and rate of return is proposed in this chapter. The Marco transform model and the AR-GARCH model are proposed for random construction in this chapter. In order to test these methods, this chapter carries out the prediction analysis within and outside the sample, and compares the accuracy of the prediction results of the two methods. The result is that the model is sufficient to capture the main characteristics such as deviation, excess kurtosis and different stages of wave rate behavior. The fourth chapter studies the fluctuation law of coal price. This chapter adopts Markoff Transform multifractal model and a series of generalized autoregressive conditional heteroscedasticity (GARCH) model to model and predict coal price volatility. This chapter extends the study [57-58] before Wei Yu et al (2010) and Wang Yudong (2016). The prediction performance of all these models is evaluated with predictive superiority test (SPA). In order to predict coal accurately The price volatility, this chapter attempts to apply a variety of different types of MSM models. This paper also considers the volatility and VaR two kinds of risk measurement methods. By comparing the models used in other research and the prediction of the MSM model, this chapter determines that the new MSM model is superior to the other models in the various prediction time domain. This superiority is also applicable to VaR. The fifth chapter studies the hedging strategy of aggregate price risk in thermal power enterprises. Using the conditional Copula function system, this chapter can model the correlation structure and the price volatility of each commodity respectively. The previous chapters have screened the most suitable model design specifications for the fluctuation rules of coal, electricity and carbon dioxide emissions quotas. In this chapter, we find a student t-Copula function which is most suitable for connecting coal, electric power MSM model and carbon dioxide emission quota GARCH model, thus establishing the Copula-MSM-GARCH model and using the EEX data of the European energy exchange. This chapter also shows how the Copula-MSM-GARCH model is used to find the smallest C with the Monte Carlo simulation method. VaR commodity combination, and proposed that the minimum CVaR portfolio can be used as a benchmark for risk hedging strategy. The adjustment of price risk combination in real production to minimum CVaR portfolio can minimize expected price risk by futures trading. The sixth chapter is the conclusion and prospect. This chapter summarizes the research results of the full text, and puts forward the research paper. Innovation and research deficiencies, and prospects for the next step.

【学位授予单位】：中国矿业大学(北京)
【学位级别】：博士
【学位授予年份】：2017
【分类号】：F224;F426.61

【参考文献】