在资产定价领域，金融经济学家们对证券市场大量的风险因子已经做出了检验。然而，对于股票风险因子的讨论非常充分，债券风险因子的研究还相对较少，中国的债券定价因子相关研究则更少。目前，我国债券市场规模已超越日本，成为亚洲第一，全球第二大债券市场。在债市市场化改革、金融监管趋严以及打破刚性兑付的大背景下，公司债务违约和债券价格大幅下跌的事件频繁发生。在此背景下，探讨债券定价和极端损失风险有助于防范与化解债券市场风险。首先，本文以债券历史收益率5%在险价值VaR作为债券极端损失的代理变量。以2012-2021年中国债券的面板数据为样本，通过单因素组合分析法、双因素组合分析法和面板回归分析，发现债券的极端损失是影响债券期望回报的重要预测指标。将极端风险特征拆分为不同的高阶矩如：波动率、偏度和峰度，研究高阶矩对债券期望收益率的影响，实证结果显示，债券收益率的波动率、偏度和峰度均对期望收益率产生显著的影响，其中波动率和峰度影响显著为正，偏度的影响显著为负。将债券按照极端损失分组，构建多空组合，并以多空组合的加权平均收益率作为债券市场下行风险因子的风险溢价。下行风险因子的风险溢价围绕1.5%左右波动，但在2020-2021年期间，波动快速上升。使用Fama-Macbeth两步法回归模型检验下行风险因子在截面上的债券定价能力，模型显示，下行风险因子在截面上对债券的超额收益有显著为正的影响。本文还对下行风险因子溢价进行了讨论，我国房地产周期是我国债券的下行风险因子溢价的重要影响因素，而债券的到期期限、净资产收益率和隐性负债率等特征对债券的下行风险暴露影响显著。本文的贡献与创新之处在于以下几个方面：第一、本文分析了单个公司的极端损失风险特征与债券预期收益率之间的关系，也从高阶矩的角度对极端损失风险的来源进行了阐述；第二、以前的学者对中国债券收益率的风险因子的研究相对较少，本文按极端损失风险特征将债券构建多空组合，作为中国债券市场的下行风险因子，并对下行风险因子进行了检验与分析；第三、本文重点讨论了下行因子风险溢价的影响因素，也讨论了债券的特征与债券下行因子风险暴露的影响因素。

Scholars have tested many risk factors in the securities markets. The discussion of the risk factors of stocks has been full, however the research of bonds’ risk factors is relatively less. At present, the scale of Chinese bond market has surpassed that of Japan, becoming the largest bond market in Asia and the second largest bond market in the world. Under heavier regulation, the defaults of corporate bond and sharp falls of bond prices have occurred frequently. In this paper, discussing bond pricing and extreme downside risk is helpful to prevent and resolve bond market risks.First, this paper uses VaR, which stands for Value at Risk, as the proxy variable of bond extreme downside risk characteristics. Through single-factor analysis and two-factor analysis, it is found that the extreme downside risk of bonds is an important predictor of the expected returns of bonds. Taking bonds’ panel data as a sample, introducing control variables and establishing a two-step regression model, the empirical results show that the impact of extreme downside risk on the expected return of bonds is still significant. Split extreme downside risk characteristics into different moments such as: volatility, skewness, and kurtosis. Using single-factor combination analysis and two-step regression, the empirical results show that the volatility, skewness, and kurtosis of bond yields all have a significant impact on the expected yield. Volatility and skewness have a significant impact is positive, and the effect of kurtosis is significantly negative.The contributions and innovations of this paper lie in the following aspects: first, previous scholars have done relatively little research on the risk factors of China‘s bond yield. Second, this paper analyzes the relationship between the extreme downside risk characteristics of a single company and the expected return rate of bonds, and also expounds the source of extreme downside risk from the perspective of high-order moments; third, this paper focuses on discussing the influencing factors of the extreme downside risk premium are discussed, and the relationship between the characteristics of bonds and the extreme downside risk exposure of bonds is discussed.