原油市场风险度量的多尺度建模理论与实证研究

Within the emerging multiscale analysis framework, this project proposes the multiscale based risk measurement theories and models, and conducts the comprehensive empirical studies in the crude oil market. The main focuses and contributions of the researches in this project involve the following aspects: 1) This project proposes the integrated statistical testing methodology for the multiscale data features in the crude oil markets. Then this project further proposes the multiscale based risk measurement theories to analyze and explain the influencing factors and transmission mechanisms for the crude oil price volatility, as well as the risk dependence structure between the crude oil market and financial markets. 2) This projects extends the traditional risk measurement models to propose the multiscale risk measurement model in the crude oil market. 3) This projects introduces the multivariate multiscale analysis to analyze the multiscale risk dependence structure and proposes the multiscale portfolio risk measurement models in the crude oil market. 4) This project conducts comprehensive empirical studies using the proposed multiscale risk analysis theory and multiscale risk measurement models. This project further provides some policy suggestions based on the results from the empirical studies. The implications of the present project is as follows: 1) Theoretically this projects proposes the theoretical framework as well as modeling methodology for multiscale risk measurement in the crude oil market, which fills the research gaps in the risk measurement and risk management fields 2) Practically results from the empirical studies using the proposed models in this project would serve as the basis for the quantitative decision making process for both enterprise operation management and government decision making process.

本项目基于多尺度分析框架，对原油市场风险度量的多尺度建模理论与实证研究进行了系统研究。主要研究内容和创新包括：1) 提出基于多尺度分析的原油市场风险特征分析框架，并在此基础上提出原油市场风险度量的多尺度分析理论,对原油市场风险波动，原油市场与经济金融系统间风险联动与投资组合风险波动进行理论分析；2）对传统风险度量模型在时频空间进行扩展，提出新的原油市场内多尺度风险度量模型；3) 基于高维多尺度分析，提出新的市场间风险联动分析与投资组合市场风险度量模型；4)利用多尺度风险分析理论及开发的多尺度风险度量技术，对原油市场风险度量进行分析，为相关部门提供决策支持。本项目的研究意义在于：1）理论上，提出多尺度风险度量理论框架与模型，解决一系列多尺度风险建模的难点问题，丰富和发展风险管理理论与技术；2) 应用上，项目的研究成果为企业生产运作和政府决策提供有益的决策依据。

该项目主要针对原油市场风险测度的多尺度建模理论与实证分析进行系统研究，在以经验模态分解、变分模态分解和多元经验模态分解等多尺度分析模型为基础的多尺度风险测度建模理论和风险预测模型构建研究中取得了一系列进展与研究成果。主要的研究结果包括：1) 基于多尺度分析理论，提出了原油市场风险测度和风险预测的多尺度建模理论与预测框架；2) 基于主流多尺度分析方法，结合计量经济模型与深度学习模型，开发了多个原油市场风险测度VaR预测模型；3) 基于多元多尺度分析方法，结合多元计量模型，建立了多个基于跨市场相依关系建模分析的投资组合风险测度VaR预测模型，并进行了相应的实证分析。截止目前，该项目已发表13篇期刊论文，其中13篇被SCI/SSCI检索；已发表3篇会议论文，其中3篇被EI Compendex检索；获得1项省级奖励，项目培养在读博士1人、毕业博士1人、毕业硕士4人。根据研究计划与完成工作，该项目已经完成项目预期目标。