Title: Identifying leverage effect in intra-day volatility pattern: Toward a functional data analysis
Authors: Ping Chen Tsai - Southern Taiwan University of Science and Technology (Taiwan) [presenting]
Abstract: Estimating intra-day volatility pattern (IVP) consists of obtaining the weights of volatility over a high dimension of intra-day intervals. The natural ordering of intra-day intervals, however, renders an interpretation of IVP as functional data. A new stylized fact of volatility is documented by identifying leverage effect in the U-shape intra-day volatility pattern, with days following negative and positive returns presenting different such patterns. Functional forms for the IVPs are then determined by non-parametric methods including interpolation and smoothing. The obtained functional forms of IVP can be incorporated into the estimation of realized variance and realized bi-power variation, which in turn has an important impact on testing for jumps in prices. A Monte Carlo study shows that the overall size and power of jump tests are improved after accounting for the leverage effect in IVP.