Authors: Svetlana Litvinova and Mervyn Silvapulle
Abstract: In risk management areas such as reinsurance, the need often arises to construct a conﬁdence interval for a quantile in the tail of the distribution. While diﬀerent methods are available for this purpose, doubts have been raised about the validity of full-sample bootstrap. In this paper, we ﬁrst obtain some general results on the validity of full sample bootstrap for the tail quantile process. This opens the possibility of developing bootstrap methods based on tail statistics. Second, we develop a bootstrap method for constructing conﬁdence intervals for high-quantiles of heavy-tailed distributions and show that it is consistent. In our simulation study, the bootstrap method for constructing conﬁdence intervals for high quantiles performed overall better than the data tilting method; the data tilting method appears to be currently the preferred choice. The applicability of the bootstrap method is illustrated using the Danish fire insurance data.