Third Year Major Oral: “The Behavior of the Bandwidth Selector Based on Cross-Validation Function in the Kernel Estimate of the Realised Spot Volatility”

Speaker: Jincheng Pang, Washington University in Saint Louis

Abstract: A kernel weighted version of the standard realised integrated volatility estimator is proposed. To choose the optimal bandwidth for kernel estimator, we propose several methods based on estimates of different prediction errors. In particular, the cross-validation technique provides unbiased estimates of the prediction error based on so-called "leave one out" estimators of the volatility. By giving the detailed proof of several important lemmas mentioned in my Minor Oral, we show asymptotic normality of both the bandwidth selector and prediction error. The choice of bandwidth is discussed and data driven selection methods proposed.

Host: Jose Figueroa-Lopez

Access Zoom Meeting (Passcode: 295079)