Parametric inference for discretely observed subordinate diffusions (Q2417988)

This paper represents an investigation in the domain of parametric inference for discretely observed ergodic subordinate diffusions. In the settings, the value of the time-changed process $X_{T}$ is observed at a discrete set of times, but the diffusion $X$ and the Lévy subordinator $T$ cannot be observed. The aim is to develop an efficient estimation method applicable for general classes of $X$ and $T$. A two step estimation procedure is developed. The first section of the article provides some background on subordinate diffusions and issues on related literature. Actually, references to previous and connected results are to be found all along the work. The authors state that ``methodologically'', the present paper improves the results from two other articles: 1) [\textit{M. Kessler} and \textit{M. Sørensen}, Bernoulli 5, No. 2, 299--314 (1999; Zbl 0980.62074)], and 2) [\textit{B. M. Bibby} and \textit{M. Sørensen}, Scand. J. Stat. 28, No. 1, 99--112 (2001; Zbl 0973.60071)]. The second section is devoted to the identification problem for subordinate diffusions and the third section to the development of the two-step procedure to estimate subordinate diffusions. Asymptotic analysis of estimators is done in the fourth section and some numerical examples are presented and discussed in the fifth one.