Second-order Regular Variation Inherited from Laplace–Stieltjes Transforms

Tiantian Mao,Lei Hua
DOI: https://doi.org/10.1080/03610926.2014.922990
2015-01-01
Abstract:Let (F) over bar be the survival function of a non negative random variable satisfying that (F) over bar is of second-order regular variation with the regular variation index - < 0 and the second-order parameter < 0. Let phi(s) denote the Laplace-Stieltjes transform of F. For (m, m + 1) and - (k, k + 1) with non negative integers m,k, if m = k, it is proved that the second-order regular variation of (F) over bar implies that of ( - 1)(k)(phi((k))(0) - phi((k))(1/ )), while for m < k, the second-order regular variation of <(F)over bar> implies that of ( - 1)(k + 1)phi((k + 1))(1/ ); for both cases, the second-order parameter can be inherited. Moreover, we prove that the converses also hold true under mild conditions. As an application, we show that the second-order regular variation of (F) over bar implies that of with regularity conditions; this result proves the converse of the result established in Mao and Hu (2013). Finally, we prove that the second-order parameter can also be easily inherited by the Laplace-Stieltjes transform of the cumulative distribution function of its reciprocals.
What problem does this paper attempt to address?