CMStatistics 2018: Start Registration
View Submission - CFE
Title: Testing the rank of cojumps in high-frequency data with market microstructure noise Authors:  Wenying Yao - Deakin University (Australia) [presenting]
Lars Winkelmann - Freie Universitaet Berlin (Germany)
Abstract: A test for the rank of a cojump matrix is proposed. The matrix consists of estimated jump sizes at specific intraday time points. High-frequency trading and the market microstructure can distort a rank test. To obtain noise-robust statistics, we use a pre-average estimator of the cojump matrix. We derive an asymptotic chi-square test for its rank. Simulation shows that the proposed test has good size and power properties. We use the new test to investigate the factor structure of cojumps at macroeconomic announcement times.