Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 9.4 — ∫₀^∞ |P(X > t) − P(Y > t)| t^{κ−1} dt ≤ κ⁻¹E|(X⁺)^κ − (Y⁺)^κ|, and (9.18) when finite

Proved
GoldieRenewal.Kesten.lemma_9_4

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

heavy-tailsmomentsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1tail-integrals

Let κ>0\kappa>0κ>0 and let XXX, YYY be real random variables on a common probability space; x+=x∨0x^+ = x\vee 0x+=x∨0.

  1. (9.17), as an inequality in [0,∞][0,\infty][0,∞]:
∫0∞∣P(X>t)−P(Y>t)∣ tκ−1 dt  ≤  1κ E∣(X+)κ−(Y+)κ∣.\int_0^\infty \big|P(X>t)-P(Y>t)\big|\,t^{\kappa-1}\,dt \;\le\; \frac1\kappa\,\mathbf E\big|(X^+)^\kappa-(Y^+)^\kappa\big| .∫0∞​​P(X>t)−P(Y>t)​tκ−1dt≤κ1​E​(X+)κ−(Y+)κ​.
  1. (9.18): if E∣(X+)κ−(Y+)κ∣<∞\mathbf E\big|(X^+)^\kappa-(Y^+)^\kappa\big|<\inftyE​(X+)κ−(Y+)κ​<∞, then t↦(P(X>t)−P(Y>t))tκ−1t\mapsto (P(X>t)-P(Y>t))t^{\kappa-1}t↦(P(X>t)−P(Y>t))tκ−1 is integrable on (0,∞)(0,\infty)(0,∞) and
∫0∞(P(X>t)−P(Y>t)) tκ−1 dt=1κ E((X+)κ−(Y+)κ).\int_0^\infty \big(P(X>t)-P(Y>t)\big)\,t^{\kappa-1}\,dt = \frac1\kappa\,\mathbf E\big((X^+)^\kappa-(Y^+)^\kappa\big).∫0∞​(P(X>t)−P(Y>t))tκ−1dt=κ1​E((X+)κ−(Y+)κ).

The lemma converts the tail-difference integrals of the implicit renewal theorem into moment expressions; it is what turns Theorem 2.3 into Corollary 2.4.

Formalization Note The paper prints (9.17) as an equality "finite or infinite". Equality fails in general (for XXX, YYY independent and identically distributed and non-degenerate the left side is 000 while the right side is positive); the paper's proof establishes, and Corollary 2.4 uses, only the inequality, which is what is stated. Both sides of (9.17) are lower Lebesgue integrals, so infinite values are allowed.

Preamble
import Mathlib
Formal statement
namespace GoldieRenewal.Kesten

open MeasureTheory ProbabilityTheory
open scoped ENNReal

/-- **Lemma 9.4** (Goldie, *Implicit renewal theory and tails of solutions of random equations*,
Ann. Appl. Probab. 1(1) (1991), p. 149). Let `κ > 0` and let `X`, `Y` be random variables on a
common probability space.

* (9.17), corrected to an inequality:
  `∫₀^∞ |P(X > t) − P(Y > t)| t^{κ−1} dt ≤ κ⁻¹ E|(X⁺)^κ − (Y⁺)^κ|`, finite or infinite.
* (9.18): when `E|(X⁺)^κ − (Y⁺)^κ| < ∞`, the function `t ↦ (P(X > t) − P(Y > t)) t^{κ−1}` is
  integrable on `(0, ∞)` and `∫₀^∞ (P(X > t) − P(Y > t)) t^{κ−1} dt = κ⁻¹ E((X⁺)^κ − (Y⁺)^κ)`.

**Formalization Note** (9.17) is printed as an equality; only `≤` holds (for `X`, `Y` i.i.d. and
non-degenerate the left side is `0` and the right side positive), and the paper's proof and its
use in Corollary 2.4 need only `≤`. Both sides of (9.17) are lower Lebesgue integrals in `[0, ∞]`,
so "finite or infinite" is literal. `x⁺ = max x 0`. -/
theorem lemma_9_4 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
    (κ : ℝ) (hκ : 0 < κ) (X Y : Ω → ℝ) (hX : Measurable X) (hY : Measurable Y) :
    (∫⁻ t in Set.Ioi (0 : ℝ),
        ENNReal.ofReal (|P.real {ω | t < X ω} - P.real {ω | t < Y ω}| * t ^ (κ - 1))
      ≤ ENNReal.ofReal κ⁻¹ *
        ∫⁻ ω, ENNReal.ofReal |(max (X ω) 0) ^ κ - (max (Y ω) 0) ^ κ| ∂P) ∧
    (Integrable (fun ω => (max (X ω) 0) ^ κ - (max (Y ω) 0) ^ κ) P →
      IntegrableOn (fun t : ℝ => (P.real {ω | t < X ω} - P.real {ω | t < Y ω}) * t ^ (κ - 1))
          (Set.Ioi 0) ∧
        ∫ t in Set.Ioi (0 : ℝ), (P.real {ω | t < X ω} - P.real {ω | t < Y ω}) * t ^ (κ - 1)
          = κ⁻¹ * ∫ ω, ((max (X ω) 0) ^ κ - (max (Y ω) 0) ^ κ) ∂P) := by sorry

end GoldieRenewal.Kesten
Source
Goldie, Implicit renewal theory and tails of solutions of random equations, Ann. Appl. Probab. 1(1):126–166 (1991), DOI 10.1214/aoap/1177005985, p. 149, Lemma 9.4, (9.17)–(9.18)
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me