限价指令簿的信息内涵研究：基于市场微观结构的视角

The organization of trading in financial markets has undergone tremendous changes at the end of the 20th century. Now most of the world's stock exchanges are either pure order-driven markets with no designated market makers or hybrid markets where market makers have to compete with a limit order book. But compared with quote-driven markets, the economic interactions in order-driven markets are more complex because the associated action and state spaces are larger and the decision dates are chosen by investors rather than being predetermined. Therefore, despite their increasing importance, the literature on order-driven markets is relatively small. Limit orders and market orders are the core engine of order-driven markets. As they have a pre-specified price, limit orders cannot always be matched with contra-side orders upon arrival. In this case, all unexecuted limit orders queue up in a limit order book. Therefore, as the most important information carrier, the limit order book (thereafter as LOB) has become the research interests of many researchers and practitioners recently. The most frequently asked questions about LOB are: What information is contained in the LOB? How do investors use this information to make their trading decisions? And what's the impact of these decisions on the efficiency of markets? To answer these questions, we investigate the information content of the LOB in the context of market microstructure. Theoretically, we contribute to the existing literature by proposing three models of an order-driven market, including a modified PIN or MRR model which incorporates the information of the LOB and therefore is more suitable for the hybrid market, a tractable dynamic model of a pure order-driven market with asymmetric information, and a new theoretical model which formally imbed dynamic limit order submission decisions in a dynamic portfolio choice problem. Empirically, with the data of Mainland China and Taiwan stock markets, we examine the LOB's contribution to price discovery, the relationship between LOB and liquidity, the private information contained in LOB, and the impact of LOB information on investors' order choice, etc. Furthermore, we compare different designs of market mechanism and elaborate on the policy implications of our findings. Our work is the first one to integrate market microstructure theory with asset allocation theory. Moreover, it extends the study of order flow to order book, and analyzes the behavior of different investors from a new aspect of order choice which has not been studied before.

近年来全球大多数交易所都已经全部或部分采用了指令驱动式交易方式。限价指令簿（LOB）作为指令驱动市场的核心及最重要的信息载体，成为理论界和实务界共同关注的焦点。究竟LOB中包含哪些重要信息？投资者如何提取这些信息进行交易决策，从而对市场效率产生怎样影响？为了回答这些问题，本课题从市场微观结构信息提炼这个角度对LOB的信息内涵进行了系统性研究，在理论上提出了基于信息不对称的纯指令驱动市场动态均衡模型，以及同时考虑投资者交易策略与最优投资组合选择的综合研究框架；在实证上对LOB与价格发现、流动性、私有信息、指令选择等进行了相关实证检验，并对不同市场机制的比较和设计进行了深入探讨。课题的创新之处体现在将原本相互割裂的资产配置理论与市场微观结构理论统一到一个新的框架下，将传统的已成交指令流研究进一步拓展到指令簿，同时从指令选择这一新的角度对不同类型投资者的行为模式进行分析，填补了相应的研究空白。

近年来全球大多数交易所都已经全部或部分采用了指令驱动式交易方式。限价指令簿（LOB）作为指令驱动市场的核心及最重要的信息载体，成为理论界和实务界共同关注的焦点。因此，本课题从市场微观结构信息提炼这个角度对LOB的信息内涵进行了系统性研究，得到的重要结果包括：（1）提出同时考虑投资者交易策略与最优投资组合选择的综合研究框架。（2）关于LOB与投资者下单行为的研究发现，台指期权市场上四类投资者的下单行为都具有显著的日内效应，四类投资者提交和取消的委托单均具有显著的对角线效应；投资者提交和取消的委托单中均存在显著的方向预测信息；投资者提交的委托单中隐含着波动率预测信息，而取消的委托单中则不包含波动率预测信息；上一笔委托单类型、买卖价差、波动率、报价中点正向变动百分比对四类投资者的下单行为均具有显著的影响。（3）关于LOB信息含量的检验发现，台指期权市场日内的流动性模式在不同时间间隔下都呈现出明显的V型；标的指数市场流动性的变动和做市商行为对台指期权市场的流动性具有显著的影响；LOB的形状指标和形状不平衡指标中包含对期权未来价格变动的预测信息，LOB的不同指标以及不同位置的信息对未来期权价格变动的预测能力存在显著差异；台指期权市场上部分个人投资者具备利用私有信息获取超额收益的能力；不同类型投资者的LOB信息对期权隐含波动率的影响机制存在显著差异。（4）关于关联市场LOB信息传导的研究发现，股票指数、股指期货和股指期权在短期内存在显著的相互引领关系，而股指期权在长期的价格调整速度最快；从全市场层面来看，台指期货自身交易的不平衡性和台指期权交易的不平衡性对未来股指收益都有显著为正的预测力，从投资者类别和期权特征分类来看，境外机构投资者、看跌期权、虚值和实值期权交易不平衡性对未来股指收益均有显著为正的预测力；知情交易者运用台指期权既交易指数的方向信息也交易指数的波动信息，但相比较而言，知情交易者更倾向于在台指期权市场交易指数的波动信息，而在台指期货市场交易指数的方向信息。（5）关于LOB与私有信息的研究发现，台指期权市场上境外和境内机构投资者都拥有私有信息，交易纪律和私有信息对两类机构投资者的绩效均有显著的影响，且交易纪律和投资者的信息优势存在交叉效应的影响。