中国债券市场流动性度量方法的比较

债券市场作为非集中化场外交易市场，如何准确地度量其流动性是金融市场微观结构领域中的关键问题。基于2008—2014年中国公司债券市场的高频逐笔交易数据，计算了4种交易成本和一种价格响应指标作为高频流动性基准指标，拓展了Corwin和Schultz的High-Low估计方法，综合考虑更多的矩条件，进一步提出了新的基于价格极值的有效价差估计，通过比较多种低频流动性指标与高频基准指标间的估计误差和相关系数，研究不同的低频流动性指标对中国债券市场的适用性。实证结果表明，对于低频交易成本指标，FHT估计总是具有最小的估计误差和最强的相关性，拓展的基于价格极值的High-Low估计和Roll估计次之；对于低频价格响应指标，Amihud指标的相关性最强。不同流动性实证研究对流动性低频指标的要求侧重不同:交易策略和资产配置以及市场有效性等相关研究侧重于低频流动性指标的估计精度；资产定价等相关问题侧重于流动性指标的相关性强弱。实证研究中可以根据对中国债券市场研究问题的侧重点不同，选择恰当的流动性度量指标。

A Comparison of Liquidity Measures on Chinese Bond Market

For the bond market, an over the counter market, how to measure the liquidity of bond market accurately is a crucial problem in market microstructure. Adopting intraday data from 2008 to 2014 on China''s corporate bond market, expanding the High-Low estimation method proposed by Corwin and Schultz,estimates of new and widely employed low-frequency liquidity measures were made in the literature against liquidity benchmarks including four trading costs and one price impact measures, and the applicability of different liquidity measures on China''s bond market was explored. The empirical results reveal that among the trading cost measures, the FHT estimator enjoys the lowest estimation error and highest correlation with benchmarks, and the extended High-Low estimator and Roll estimator take the second place, while Amihud liquidity ratio measure owns the highest correlation through all price impact measures. Since the related issues such as asset pricing focus on correlation performance of the low frequency liquidity measures, and liquidity estimation precision is mostly associated with market effectiveness test and corporate finance, the appropriate liquidity measures should be adopted due to different research emphasis of the bond market''s empirical studies.