A0913
Title: State-space method for the quadratic estimator of integrated variance in the presence of market microstructure noise
Authors: Daisuke Nagakura - Keio University (Japan) [presenting]
Toshiaki Watanabe - Hitotsubashi University (Japan)
Abstract: Recently, a class of integrated variance (IV) estimators, called the quadratic estimator (QE), has been considered which takes a quadratic form in observed returns. The QE includes several existing IV estimators as special cases, such as the realized variance, two-time scale estimator, and realized kernels. Even in the presence of market microstructure noises (MMNs) in observed prices, some special cases of the QE are consistent. However, they still have finite sample biases due to MMNs. We derive a multivariate state-space representation of QEs, where the IV is a common component in these QEs. Applying the Kalman filter to the state-space representation provides the linear projection of IV on these QEs which has the smallest MSE of any QEs employed in the state space representation. We conduct some simulation studies to check the performance of our method, and then do some empirical analysis, applying our method to the actual data.
