Title: Transition and measurement noise correlation in affine and Gaussian models: the case of oil prices
Authors: Carla Gomes Costa De Souza; Fernando Antonio Lucena Aiube
Addresses: Faculty of Economic Sciences, State University of Rio de Janeiro, Rua São Francisco Xavier, 524, Maracanã, Rio de Janeiro, 20550-900, Brazil ' Faculty of Economic Sciences, State University of Rio de Janeiro, Rua São Francisco Xavier, 524, Maracanã, Rio de Janeiro, 20550-900, Brazil; Graduate Program in Administration, Fluminense Federal University, Rua Mário Santos Braga, Niterói, Rio de Janeiro, 24.020-140, Brazil
Abstract: This paper proposes a new approach for the estimation of affine and Gaussian factor models with the Kalman filter method. It considers the correlation between the innovations of transition and measurement equations. We use crude oil prices in the analysis. When applying this correlation approach in two- and three-factor models, we obtain improvements of error measures between estimated and observed future prices with inexpensive estimation procedures.
Keywords: commodity factor models; Kalman filter estimation; future prices; oil prices.
International Journal of Financial Markets and Derivatives, 2021 Vol.8 No.1, pp.50 - 64
Received: 12 Jun 2020
Accepted: 18 Aug 2020
Published online: 31 Mar 2021 *