“…In the nonparametric case, let X 0 , X 1 , X 2 be independent random variables with distribution functions G, F 1 and F 2 , density functions g, f 1 and f 2 , survival functionsḠ,F 1 and F 2 and failure rate functions r G , r F 1 and r F 2 , respectively (note r G (x) = g(x)/Ḡ(x).) Then Bagai and Prakasa Rao (1992) showed that when T, δ are observable, then in the class of distributions with X 0 fixed, individual distributions F 1 and F 2 are not identifiable but the ratio of failure rates H defined by…”