Title: Jump testing with the pre-averaged bipower variation and subsampling estimation of the asymptotic variance matrix
Authors: Bezirgen Veliyev - Aarhus University (Denmark) [presenting]
Kim Christensen - Aarhus University (Denmark)
Nopporn Thamrongrat - Heidelberg University (Germany)
Abstract: A noise-robust extension of the bipower variation-based jump test is proposed, which is based on pre-averaging and can implemented on high-frequency data that are perturbed by microstructure noise. The main hurdle in this context is to derive a consistent estimator of the asymptotic covariance matrix of the pre-averaged bipower variation, which is asymptotically jump-robust and has good finite sample properties, such as being positive semi-definite and well-conditioned. We propose such an estimator based on a subsampling approach. Simulations show that the proposed test has size control under a variety of noise structures, while it has excellent power under the alternative. In the empirical application, we recover jump dates from real data.