General regular variation of \(n\)\,th order and the 2nd order Edgeworth expansion of the extreme value distribution. II (Q2431884)

In part one of this paper [ibid. 21, No. 5, 1121-1130 (2005; Zbl 1081.62009)] the concept of general regular variation of \(n\)\,th order has been introduced. In this second part the authors use this concept to obtain a second order Edgeworth expansion for shifted and scaled sample maxima \(a_n^{-1} (\max \{ X_1, \ldots, X_n\} - b_n)\) of i.i.d.~sequences \((X_n)_{n\geq 1}\). This expansion is obtained under the assumption that the left continuous inverse of \(-1/\log F\) is general regularly varying of third order, where \(F\) denotes the distribution function of \(X_1\). Some inequalities for general regularly varying functions of third order are also obtained.