Detecting factors of quadratic variation in the presence of market microstructure noise

Abstract A method of detecting latent factors quadratic variation (QV) Itô semimartingales from a set discrete observations is developed when the market microstructure noise present. We propose new way to determine number co-variations asset prices based on SIML (separating information maximum likelihood) by Kunitomo et al. (Separating likelihood estimation for high frequency financial data. Springer, Berlin, 2018). In high-frequency data, it important investigate effects possible jumps and existed in markets. explore estimated variance–covariance matrix (efficient) underlying its characteristic roots vectors variation. give some simulation results see finite sample properties proposed illustrate an empirical data analysis Tokyo stock market.