§2.1, p. 5 — existence of uniform on with a.s. on an atomless space
ProvedTailRiskSharing.VaRTail.exists_quantile_uniformatomlessquantile-functiontail-risk-measureuniform-random-variable
Let be an atomless probability space (every measurable set with contains a measurable with ). Then for every random variable there exists a random variable , uniformly distributed on , such that
where is the left-continuous quantile function of .
In the formalization, "uniform on " means is measurable with law , and is expressed as (the predicate IsQuantileUniform). This is the existence statement that §2.1 of the paper invokes for every and that Theorem 2 (iv) relies on to compare across tail-equivalent random variables.
Preamble
import Mathlib import Definitions.Def_TailRiskSharing_VaRTail_Setting open MeasureTheory
Formal statement
namespace TailRiskSharing.VaRTail
theorem exists_quantile_uniform {Ω : Type*} [MeasurableSpace Ω]
(P : Measure Ω) [IsProbabilityMeasure P] (hP : TailRiskSharing.VaRConv.IsAtomless P)
(X : Ω → ℝ) (hX : X ∈ (TailRiskSharing.VaRConv.L0 : Set (Ω → ℝ))) :
∃ U : Ω → ℝ, IsQuantileUniform P X U := by sorry
end TailRiskSharing.VaRTailSource
Liu, Mao, Wang & Wei, Inf-convolution, Optimal Allocations, and Model Uncertainty for Tail Risk Measures, SSRN 3490348 (version of January 16, 2022), §2.1 p. 5 (existence of U_X, citing Föllmer & Schied, Stochastic Finance, 4th ed. 2016, Lemma A.32)