中国国债市场收益率曲线套利策略

发布时间：2018-04-10 15:38

本文选题：收益率曲线 + 套利策略　；参考：《上海交通大学》2013年硕士论文

【摘要】：私募基金的快速成长使得市场对高效、高收益的对冲策略需求日益旺盛，而中国债券市场的迅速扩容为固定收益对冲策略的开发提供了充分的产品和平台支持。实践中虽有少数投资公司在逐步探索收益率曲线套利策略，但学术上仍缺乏对该策略详尽和坚实的验证和风险评估。基于此，本文试图提供一个包括收益率曲线套利策略构建、风险因素分析及收益评价的综合分析框架。此外，基于两因素Vasicek模型，本文还考察了中国国债市场的定价误差。本文选取了2006年7月至2012年6月共六年的国债数据进行实证。样本包括了银行间及上交所所有国债。样本区间及市场的选取是基于交易量大小（流动性）及发展情况而定。在两因素Vasicek模型的框架下，本文完整的推导了国债理论价格公式。基于此，通过一定的规则确定了最优套保比率和无套利空间并且计算出涵盖买卖价差、佣金、冲击成本及回购成本在内的交易成本，根据一定的交易规则计算出了8个子策略超额收益率并发现策略能够获得显著正超额收益，呈尖峰右侧厚尾分布，波动小、组合调整时间小于6个月等特点。本文进一步展开对8个收益率序列的分析，发现交易成本是超额收益的重要解释变量，而无套利空间的变化虽然对风险调整收益无明显影响，但却显著改变了收益的概率分布。利用资本资产定价模型（CAPM）和Fama-French多因素模型，，本文考察了策略收益的风险因子。我们发现CAPM不能很好的解释超额收益，两个模型对基于交易所市场的策略解释效果尤其不佳，即便在添加了流动性风险因子的情况下。策略收益与Fama-French因子呈非线性关系。此外，利用多因素模型，我们发现策略失败概率与流动性强度负相关、与宏观经济走势正相关。基于传统度量指标和纳入非线性和非正态分布的有效性检验，我们发现收益率曲线套利策略能够在短期和长期中产生显著正超额收益，但并不能提供更具优势的投资收益。
[Abstract]:The rapid growth of private equity funds makes the demand for efficient and high-yield hedging strategies increasingly strong, while the rapid expansion of China's bond market provides sufficient product and platform support for the development of fixed-income hedging strategies.Although a small number of investment companies are gradually exploring the arbitrage strategy of the yield curve in practice, there is still a lack of detailed and solid verification and risk assessment of the strategy.Based on this, this paper attempts to provide a comprehensive analysis framework including the arbitrage strategy of yield curve, risk factor analysis and income evaluation.In addition, based on two-factor Vasicek model, the pricing error of Chinese government bond market is investigated.This paper selects the national debt data from July 2006 to June 2012 for empirical analysis.The sample includes all inter-bank and SSE Treasuries.Sample ranges and markets are selected based on volume size (liquidity) and developments.Under the framework of two factors Vasicek model, the theoretical price formula of treasury bonds is derived in this paper.Based on this, the optimal arbitrage ratio and no arbitrage space are determined by certain rules, and the transaction costs, including purchase and sale spread, commission, impact cost and repo cost, are calculated.According to certain trading rules, the excess return rate of 8 sub-strategies is calculated and it is found that the strategy can obtain significantly positive excess returns, with the characteristics of a thick tail distribution on the right side of a spike, a small fluctuation, and a combination adjustment time of less than 6 months.This paper further analyzes the eight return sequences and finds that transaction cost is an important explanatory variable of excess return, while the change of no arbitrage space has no obvious influence on risk adjusted return, but it changes the probability distribution of income significantly.Using the capital asset pricing model (CAPM) and Fama-French multi-factor model, this paper investigates the risk factors of strategic returns.We find that CAPM can not explain the excess return very well, and the two models are especially bad for the strategy explanation based on the exchange market, even if the liquidity risk factor is added.There is a nonlinear relationship between the policy return and the Fama-French factor.In addition, by using multi-factor model, we find that the probability of strategy failure is negatively correlated with liquidity intensity and positively correlated with macroeconomic trend.Based on the traditional metrics and the validity test including nonlinear and non-normal distribution, we find that the yield curve arbitrage strategy can produce significantly positive excess returns in the short and long term, but can not provide more advantageous investment returns.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F812.5;F832.51

【参考文献】