An estimator of the ex-post covariation of log-prices under asynchronicity and microstructure noise is proposed. It uses the Cholesky factorization on the correlation matrix in order to exploit the heterogeneity in trading intensity to estimate the different parameters sequentially with as many observations as possible. The estimator is guaranteed positive semidefinite. Monte Carlo simulations confirm good finite sample properties. In the application we forecast portfolio Value-at-Risk and sector risk exposures for a portfolio of 52 stocks. We find that forecasts obtained from dynamic models utilizing the proposed high-frequency estimator provide statistically and economically superior forecasts to models using daily returns.
In the present paper, we study selfdecomposability of random fields, as defined directly rather than in terms of finite-dimensional distributions. The main tools in our analysis are the master Lévy measure and the associated Lévy-Itô representation. We give the dilation criterion for selfdecomposability analogous to the classical one. Next, we give necessary and sufficient conditions (in terms of the kernel function) for a Volterra field driven by a Lévy basis to be selfdecomposable. In this context, we also study the so-called Urbanik classes of random fields. We follow this with the study of existence and selfdecomposability of integrated Volterra fields. Finally, we introduce infinitely divisible field-valued Lévy processes, give the Lévy-Itô representation associated with them and study stochastic integration with respect to such processes. We provide examples in the form of Lévy semistationary processes with a Gamma kernel and Ornstein-Uhlenbeck processes.
This paper studies some probabilistic properties of a Lévy semistationary process when the kernel is given by '˛; .s/ D e s s˛for˛> 1 and > 0. We study the stationary distribution induced by this process. In particular, we show that this distribution is self-decomposable for 1 <˛< 0 and under certain conditions it can be characterized by the so-called cancellation property.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.