【作者】 刘艳萍；

【导师】 颜波；

【作者基本信息】 华南理工大学 ， 应用经济学， 2024， 博士

【摘要】 随着全球气候问题的关注度不断提高,以排放权为标的进行交易和投资的碳金融活动蓬勃发展。碳市场拥有的商品和金融双重属性,使其在经济社会发展中发挥重要作用,同时与能源、股票市场产生紧密关联。2008年以来,金融危机、英国脱欧、新冠疫情、俄乌冲突等重大事件引发的经济不确定性在全球不断蔓延,碳、能源、股票等市场遭受巨大冲击。在此背景下,各国频繁出台应对政策,使得经济政策不确定性增大,碳市场与能源、股票市场间的内在联系也随之倍加复杂。现有文献已广泛关注碳市场与关联市场之间的互动关系,但考虑经济政策不确定性(EPU)冲击作用的研究还存在提升空间。本文的研究目的是探讨当前“黑天鹅”事件频发的局势下欧盟碳市场与关联市场的溢出效应。在理论层面,搭建“EPU—碳—能源”和“EPU—碳—股票”的分析框架,梳理成本调整、政策引导、风险管理、情绪驱动四条信息传导路径,以及收益率、波动率和尾部收益率三个维度的溢出效应机理。在实证层面,首先构建计量模型探讨EPU在碳市场、能源市场和股票市场的信息传导作用,然后分别从均值溢出效应、波动溢出效应和极端风险溢出效应三个维度探讨EPU冲击下碳市场与关联市场之间的信息和风险传导机理。本文的研究不仅为政策制定者和市场投资者在宏观经济政策动荡情形下的决策提供指引,还可以为中国碳市场的建设提供参考。主要结论概括如下:第一,经济政策不确定性在各个市场的信息传导作用存在异质性。本文为解决EPU指数和交易数据频率不一致问题,应用非对称广义自回归条件异方差—混频数据抽样模型(简称:混频模型)对市场波动率进行分解,将EPU的作用反映到各个市场的长期波动率上。实证结果表明,EPU对碳市场和能源市场的长期波动有负向作用,对股票市场的长期波动有正向作用,且对能源、碳、股票市场的作用强度依次减小。不同水平的EPU、不同性质的重大事件所引发的EPU、EPU的非对称冲击对各市场长期波动的影响存在差异。此外,模型置信集合检验结果证实,同时考虑已实现波动率和EPU的双因子混频模型的样本外预测性能最佳。第二,投资者情绪在经济政策不确定性和收益波动中发挥桥梁作用。碳市场与关联市场存在正向为主的非对称相关关系,且相关系数的波动趋势与重大事件有关。提取碳市场和关联市场的交易数据所构建的投资者情绪(EUIS)指数可以很好地综合反映碳市场投资者的心理预期和经济行为。混频模型和动态相关性模型估计结果表明,经济政策不确定性增大将缓解EUIS指数波动,而EUIS指数与碳价格波动存在负向为主的相关关系。通过构建时变脉冲响应函数发现,各个市场对EPU冲击的响应强度在短期最大,而EUIS指数的响应则由于投资者避险心理的影响在中期更加显著;各市场和EUIS对不同事件冲击的响应具有时变性和异质性。第三,不同层面波动溢出效应的表现反映长短期波动率包含信息的差异。EPU对碳市场与关联市场之间的动态相关性有显著作用。鉴于市场间的相关性变化将影响跨市场的波动信息传导,应用时变溢出指数模型分析不同波动率层面的溢出效应。静态结果表明,系统在长期波动层面的溢出效应大于总波动和短期波动,且正向EPU冲击下的长期波动溢出强度最大;市场间存在双向和不对称的波动溢出,且各市场的角色在不同波动层面上有差别;经济下行时能源和股票市场的溢出强度比碳市场大,而经济形势改善会驱动碳市场的信息传导。动态结果表明,长期波动反映宏观经济形势长期均衡信号,溢出指数趋势较为平缓;短期波动受突发事件、情绪等短期事件干扰,溢出指数波动频繁;总波动包含了长短期波动的共同信息,其溢出指数的波动趋势居于二者之间。第四,经济政策不确定性的冲击作用在风险测度和资产配置中不可忽视。碳市场和能源市场的在险价值高于股票市场。碳市场与关联市场的尾部紧密相关,存在双向和非对称风险溢出效应。能源市场和股票市场是碳市场的风险源,且股票市场对碳市场的风险溢出强度更大。受重大事件的驱动,碳市场与关联市场之间的风险传递时变特征明显。在研究极端风险时,考虑EPU的冲击大小及方向可以避免各市场的在险价值及跨市场风险溢出强度的测度偏误。此外,碳期货与关联资产进行组合投资可降低风险,且最优组合权重的设置与资产类别、经济形势有关。与现有研究相比,本文在以下三个方面取得了新进展:第一,将经济政策不确定性纳入研究体系并从多个维度探讨碳市场与关联市场的溢出效应,相较以往基于稳态形势的研究更加贴近当前动荡的社会环境;第二,量化碳市场的投资者情绪,并探讨经济政策不确定性冲击下其在跨市场信息传导中的重要作用,将行为金融理论拓展到碳市场的研究中;第三,将EPU指数作为外生冲击因子,从混频模型的长短期信息分解视角开展碳市场与关联市场溢出效应的分析,避免直接将该指数作为系统内部变量进行模型估计可能引起的结果偏误。更多还原

【Abstract】 With the increasing attention to global climate issues,carbon finance activities that trade and invest based on carbon emission rights are flourishing.The dual attributes of commodities and finance in the carbon market make it play an important role in economic and social development,while closely related to energy and stock markets.Since 2008,economic uncertainty caused by major events such as the financial crisis,Brexit,COVID-19,and the Russia-Ukraine conflict has spread around the world,and the carbon,energy,stock,and other markets have suffered a huge impact.In this context,countries frequently introduce response policies,leading to increased economic policy uncertainty,and the inherent connections between carbon and energy markets,as well as carbon and stock markets,have become even more complex.The existing literature has extensively focused on the interaction between carbon markets and related markets,but there is still room for improvement in research considering the impact of economic policy uncertainty(EPU).The purpose of this article is to explore the spillover effects of the EU carbon market and related markets in the current situation of frequent occurrence of "black swan" events.At the theoretical level,this article constructs an analytical framework for "EPU-carbon-energy" and "EPU-carbon-stock",and sorts out four information transmission paths: cost adjustment,policy guidance,risk management,and emotion driven,as well as the three spillover effect mechanisms of returns,volatility,and tail returns.At the empirical level,this article first constructs an econometric model to explore the information transmission role of EPU in the carbon,energy,and stock markets.Then,the information and risk transmission mechanisms between the carbon market and related markets under the impact of EPU are explored from three dimensions: mean spillover effect,volatility spillover effect,and extreme risk spillover effect.This study not only provides guidance for policymakers and market investors in the face of drastic macroeconomic and policy fluctuations but also provides a reference for the construction of China’s carbon market.The main conclusions are summarized as follows:Firstly,there is heterogeneity in the information transmission effect of economic policy uncertainty in various markets.To address the issue of inconsistent frequency between EPU and trading data,the asymmetric generalized autoregressive conditional heteroscedasticity-mixed data sampling model(referred to as the mixing model)is applied to decompose market volatility and reflect the role of EPU on the long-term volatility of each market.The empirical results indicate that EPU has a negative effect on the long-term volatility of the carbon and energy markets,a positive effect on the long-term volatility of the stock market,and the intensity of its effect on the energy,carbon,and stock markets decreases in sequence.Different levels of EPU,EPU triggered by major events of different natures,and asymmetric shocks of EPU have heterogeneous effects on the volatility of various markets.In addition,the model confidence set test confirms that the out-of-sample forecasting performance of the mixing model considering both realized volatility and EPU is the best.Secondly,investor sentiment plays a bridging role in economic policy uncertainty and returns volatility.There is a positive asymmetric correlation between the carbon market and related markets,and the fluctuation trend of the correlation coefficient is related to major events.The investor sentiment index(EUIS)constructed by extracting trading data from the carbon market and related markets can effectively reflect the psychological expectations and economic behavior of carbon market investors.The mixing model and dynamic correlation analysis indicate that an increase in economic policy uncertainty will alleviate the volatility of the EUIS index,and there is a negative correlation between the EUIS index and carbon price fluctuations.By constructing time-varying impulse response functions,it is found that the response intensity of each market to EPU shocks is the highest in the short term,while the response of the EUIS index is more significant in the medium term due to the influence of investor hedging psychology;The response of each market and EUIS to different events is time-varying and heterogeneous.Thirdly,the performance of spillover effects at different volatility levels reflects the differences in information contained in long-term and short-term volatilities.EPU has a significant impact on the dynamic correlation between the carbon market and related markets.Given that changes in correlation between markets will affect the transmission of volatility information across markets,a time-varying spillover index model is applied to analyze the spillover effects at different volatility levels.The static results indicate that the total spillover effect of the system at the long-term volatility is greater than that of the total volatility and short-term volatility,and the spillover intensity of the long-term volatility is the highest under positive EPU shocks;There are bidirectional and asymmetric spillover effects between markets,and the roles of each market vary at different volatility levels;During economic downturns,the spillover intensity of energy and stock markets is greater than that of the carbon market,and improving economic conditions will drive information transmission in the carbon market.The dynamic results suggest that the long-term volatility reflects the long-term equilibrium signal of the macroeconomic situation,with a smoother trend;The short-term volatility is affected by short-term events such as unexpected events and sentiment,with relatively frequent fluctuations;The total volatility contains the common information of long-term and short-term volatilities,and its spillover index trend falls between the two.Fourthly,the impact of economic policy uncertainty cannot be ignored in risk measurement and asset allocation.Individual risks in carbon and energy markets are higher than those in stock markets.The carbon market is closely related to the tail returns of related markets,and there is a bidirectional and asymmetric risk spillover between the carbon market and related markets.Energy and stock markets are the sources of risk to the carbon market,and the risk spillover intensity of stock markets to the carbon market is greater than that of energy markets.Driven by major events,the time-varying characteristics of risk transmission between the carbon market and related markets are evident.When studying extreme risks,considering the magnitude and direction of EPU shocks can avoid measurement bias in the value at risk of each market and cross-market risk spillover.In addition,combining carbon futures with related assets can reduce risk,and the setting of optimal portfolio weights is related to asset categories and economic conditions.Compared with existing research,this article has made improvements in the following three aspects.Firstly,this article incorporates economic policy uncertainty into the research system and explores the spillover effects of the carbon market and related markets from multiple dimensions,which is closer to the current turbulent social environment than previous studies based on steady-state situations;Secondly,this article quantifies investor sentiment in the carbon market and explores its important role in cross-market information transmission under the impact of economic policy uncertainty,which extends behavioral finance theory to the study of carbon markets;Thirdly,this article takes the EPU index as an exogenous shock factor and analyzes the spillover effects of the carbon market and related markets from the perspective of the long-term and short-term information decomposition by the mixed frequency model,avoiding the possible result bias of directly using the index as an internal variable of the system for model estimation.更多还原