Leverage effect in high-frequency data with market microstructure

The leverage effect is an important explanation for volatility asymmetry, which has got extensively attention in the recent years. In this paper, we introduces a new estimator of leverage effect. The key feature of the proposed estimator is explored in the setting when the microstructure noise model is the parameter function of trading information. The proposed estimator shows good statistical performances via theorems and simulations study. Specially, the estimator has a convergence rate n^1/4. The QQ-Plots, Histogram plots and quartiles perform sufficient asymptotical normality compared with the exist estimated methods. An empirical study is carried out to demonstrate that the proposed estimator could present the efficient application value, and confirm that the leverage effect plays an important role in forecasting volatility.