Autoregressive inverse Gaussian process and the stochastic volatility modeling. Issue 10 (19th May 2023)

• Record Type: Journal Article
• Authors: Sujith, P; Balakrishna, N.
• Is Part Of: Communications in statistics. Volume 52:Issue 10(2023)
• Page Start: 3574
• Page End: 3580
• Publication Date: 2023-05-19
• Subjects: Normal inverse Gaussian -- Inverse Gaussian -- autoregressive -- stochastic volatility -- Levy measure; Mathematical statistics -- Periodicals; Mathematics; Statistics; 519.2
• DOI: 10.1080/03610926.2021.1977324
• Languages: English
• ISSNs: 0361-0926

Abstract

Abstract: Normal mixture of inverse Gaussian known as Normal-inverse Gaussian distribution is well-known in the context of modeling the stochastic volatility. In the present work, an autoregressive model for generating the sequence of volatilities is suggested so that the return series will have a stationary normal-inverse Gaussian distribution. The Laplace Transform of the innovation distribution does not have a closed form for its inversion. So the distributional properties are studied in terms of the Levy measure. The model parameters are estimated by the method of moments and a simulation is carried out to check their performance.