Bounds for the sum of dependent risks and worst value-at-risk with monotone marginal densities (Q1945047)

The main result in this paper establishes sharp dependence bounds for the distribution function of a sum of dependent, identically distributed risks with monotone density. As a consequence this result implies sharp bounds for the value at risk. The proof of this result is based on mixability properties of the distributions. It extends a previous result of \textit{B. Wang} and the first author [J. Multivariate Anal. 102, No. 10, 1344--1360 (2011; Zbl 1229.60019)] which was concerned with the case of random variables in \([0,1]\).