On the Numerics of Estimating Generalized HyperbolicDistributions

Abstract

An important empirical fact in financial market is that return distributions are often skewed and heavy-tailed. This paper employs maximum likelihood estimation to estimate the five parameters of generalized hyperbolic distribution, a highly flexible heavy-tailed distribution. The estimation utilizes Powell’s methods in multidimensions and the performance of estimation is measured by simulation studies. Application to the financial market provides us with estimates of return distribution of some financial assets.