The log-L\'evy moment problem via Berg-Urbanik semigroups

TL;DR

This paper investigates the transition point between moment determinacy and indeterminacy for Berg-Urbanik semigroups, linking it to properties of the underlying Bernstein function, and provides detailed distributional analysis.

Contribution

It extends previous work by estimating the threshold time for determinacy transition using properties of Bernstein functions and introduces new distributional results for these semigroups.

Findings

Identifies the threshold time for moment indeterminacy transition.

Provides detailed asymptotic behavior of the semigroup density.

Introduces a non-classical Abelian criterion for the moment problem.

Abstract

We consider the Stieltjes moment problem for the Berg-Urbanik semigroups which form a class of multiplicative convolution semigroups on $\mathbb{R}_+$ that is in bijection with the set of Bernstein functions. Berg and Dur\'an proved that the law of such semigroups is moment determinate (at least) up to time $t=2$, and, for the Bernstein function $\phi(u)=u$, Berg made the striking observation that for time $t>2$ the law of this semigroup is moment indeterminate. We extend these works by estimating the threshold time $\scr{T}_\phi \in [2,\infty]$ that it takes for the law of such Berg-Urbanik semigroups to transition from moment determinacy to moment indeterminacy in terms of simple properties of the underlying Bernstein function $\phi$, such as its Blumenthal-Getoor index. One of the several strategies we implement to deal with the different cases relies on a non-classical Abelian type criterion for the moment problem, recently proved by the authors. To implement this approach we provide detailed information regarding distributional properties of the semigroup such as existence and smoothness of a density, and, the large asymptotic behavior for all $t > 0$ of this density along with its successive derivatives. In particular, these results, which are original in the L\'evy processes literature, may be of independent interests.