带有市场交易信息和随机微观噪声下的杠杆效应研究

在风险管理中杠杆效应的现象广泛存在，也是金融计量学中的重要议题。高频金融市场中蕴含着丰富的交易信息，而这些信息并不能都看作随机噪声，因此探讨利用市场交易信息并在带有随机噪声模型下研究杠杆效应具有重要意义。本文在带有市场交易信息和随机微观噪声相结合的模型下研究了杠杆效应，提出了新的杠杆效应估计，该估计具有n^1/8的收敛速度，同时给出了估计的方差和相关的定理。通过模拟分析得出利用广泛的市场微观信息可以更有效和更精确地对杠杆效应进行估计，模拟的结果表明本文提出的杠杆效应估计具有更好的渐近正态性和更小的偏差。最后将提出的估计应用到实证分析中，发现杠杆效应对未来一天波动率的预测具有显著性影响。

Leverage Effect Combining Trading Information with Stochastic Microstructure Noise

Leverage effect is a widespread existence in risk management, which is also an important topic in financial econometrics. On the other hand, it is valuable for us to study leverage effect based on the market trade information and the stochastic noise model, since that rich information is available and not viewed as the stochastic noise in high-frequency financial market. In this paper, the leverage effect combining trading information with stochastic microstructure noise model is studied, and the new estimator of leverage effect is proposed, which has the convergence rate n^1/8. The variance and the relative theorems are also provided. A simulation is conducted to study that leverage effect could be estimated more efficiently and more accurately via the general market microstructure information. Simulation results show the leverage effect estimator performs better asymptotic normal property and smaller biases. An empirical study has been carried out to demonstrate that leverage effect is significant in forecasting volatility, which provides a useful theoretical basis for relevant financial departments to conduct financial guidance.