On the moments of the variance-gamma distribution (Q6178685): Difference between revisions

The author obtains new closed-form formulas for the moments and absolute moments of the variance-gamma distribution in the following main theorem: Theorem: Let $X\sim VG(\nu,\alpha,\beta,0)$, where $\nu>-1/2$, $0\leq|\beta|<\alpha$. Then, for $k>k_*$, \[ E(|X|^{k})=((2^{k}(1-\beta^2/\alpha^2)^{\nu+1/2})/(\sqrt{\pi\alpha}^{k}\Gamma(\nu+1/2)))\] \[ \Gamma(\nu+((k+1)/2))\Gamma(((k+1)/2))_2F_1(((k+1)/2),\nu+((k+1)/2);(1/2); ((\beta^2)/(\alpha^2))), \] and, for odd $k\in Z^+$ \[ E(X^{k})=((2^{k+1}\beta(1-\beta^2/\alpha^2)^{\nu+1/2})/(\sqrt{\pi\alpha}^{k+1}\Gamma(\nu+1/2)))\] \[ \Gamma(\nu+(k/2)+1)\Gamma((k/2)+1)_2F_1((k/2)+1,\nu+(k/2)+1;(3/2);((\beta^2)/(\alpha^2))), \] where $VG(\nu,\alpha,\beta,\mu=0)$ is the variance-gamma distribution with parameters $\nu>-1/2$, $0\leq |\beta|<\alpha$, $\mu\in R$; $k_*=k_*(\nu):=\max\{-1,-2\nu-1\}$ and $_2F_1$ is a hypergeometric function [\textit{F. W. J. Olver} (ed.) et al., NIST handbook of mathematical functions. Cambridge: Cambridge University Press (2010; Zbl 1198.00002)]. He deduces new formulas for the moments and absolute moments of the product of two correlated zero mean normal random variables in some remarks.