Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 1, p. 12 — optimal profit and optimal quantity are higher under partial refunds than under full refunds

Proved
SuReturns.PartialRefunds.corollary1_partial_beats_full

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

consumer-returnsnewsvendorp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1partial-refund

Standing assumptions (Su pp. 6–7): market demand X∼DX \sim DX∼D with X≥0X \ge 0X≥0; consumer valuations V∼νV \sim \nuV∼ν with finite mean μ=EV\mu = EVμ=EV; unit cost ccc and salvage value sss with 0≤s<c<μ0 \le s < c < \mu0≤s<c<μ. Let FFF be the distribution function of DDD, Fˉ=1−F\bar F = 1 - FFˉ=1−F, Gˉ(p)=P(V≥p)\bar G(p) = \mathbb P(V \ge p)Gˉ(p)=P(V≥p), Π\PiΠ the seller's profit (7) under demand rule (6), and Πfull\Pi_{\text{full}}Πfull​ the full-refund profit (1). Let (pF∗,qF∗)(p^*_F, q^*_F)(pF∗​,qF∗​) be the full-refund policy of Proposition 1(i): pF∗p^*_FpF∗​ maximizes (p−s)Gˉ(p)(p - s)\bar G(p)(p−s)Gˉ(p) over all real ppp and Fˉ(qF∗)=(c−s)/((pF∗−s)Gˉ(pF∗))\bar F(q^*_F) = (c-s)/((p^*_F - s)\bar G(p^*_F))Fˉ(qF∗​)=(c−s)/((pF∗​−s)Gˉ(pF∗​)). Let qP∗q^*_PqP∗​ be the quantity of Proposition 2: Fˉ(qP∗)=(c−s)/(Emax⁡(V,s)−s)\bar F(q^*_P) = (c-s)/(E\max(V,s) - s)Fˉ(qP∗​)=(c−s)/(Emax(V,s)−s). Then

  1. for every price ppp and every q≥0q \ge 0q≥0, Πfull(p,q)≤Π(Emax⁡(V,s), qP∗, s)\Pi_{\text{full}}(p, q) \le \Pi\big(E\max(V,s),\, q^*_P,\, s\big)Πfull​(p,q)≤Π(Emax(V,s),qP∗​,s), so the optimal partial-refund profit is at least the profit of every full-refund policy;
  2. if FFF is strictly increasing on [0,∞)[0,\infty)[0,∞), then
qF∗≤qP∗.q^*_F \le q^*_P.qF∗​≤qP∗​.

This is Corollary 1: partial refunds raise both the seller's expected profit and the stocking level.

Formalization Note Part 2 assumes FFF strictly increasing on [0,∞)[0,\infty)[0,∞); the page's proof inverts Fˉ\bar FFˉ (it writes Fˉ−1\bar F^{-1}Fˉ−1 on p. 27), which presumes this. Without it part 2 is false: on a flat stretch of FFF the two fractile equations can have the same solution set and any qF∗>qP∗q^*_F > q^*_PqF∗​>qP∗​ from it satisfies both. "Higher" is read as "at least as high", as in the page's proof (ΠP∗≥ΠF∗\Pi^*_P \ge \Pi^*_FΠP∗​≥ΠF∗​, qP∗≥qF∗q^*_P \ge q^*_FqP∗​≥qF∗​). Demand is assumed nonnegative (D((−∞,0))=0D((-\infty,0)) = 0D((−∞,0))=0) and stocking quantities range over q≥0q \ge 0q≥0: the page's demand is "a mass of infinitesimal consumers" and qqq a "stocking quantity", and without these readings the profit is unbounded (a negative qqq with a very negative price makes it arbitrarily large). 0≤s0 \le s0≤s is the page's own "0≤s≤c0 \le s \le c0≤s≤c" (p. 26). The valuation law is a general probability measure with integrable identity rather than a density ggg; Gˉ(r)=P(V≥r)\bar G(r) = \mathbb P(V \ge r)Gˉ(r)=P(V≥r) follows the page's tie rule (keep iff V≥rV \ge rV≥r).

Preamble
import Mathlib
import Definitions.Def_SuReturns_PartialRefunds_Model

open MeasureTheory ProbabilityTheory
Formal statement
namespace SuReturns.PartialRefunds

theorem corollary1_partial_beats_full (ν D : Measure ℝ) [IsProbabilityMeasure ν]
    [IsProbabilityMeasure D] (hν : Integrable (fun v => v) ν) (hD0 : D (Set.Iio 0) = 0)
    (c s : ℝ) (hs0 : 0 ≤ s) (hsc : s < c) (hcμ : c < meanValuation ν)
    (pF qF : ℝ) (hp : IsMaxOn (fun p => (p - s) * keepProb ν p) Set.univ pF)
    (hqF : 1 - cdf D qF = (c - s) / ((pF - s) * keepProb ν pF))
    (qP : ℝ) (hqP : 1 - cdf D qP = (c - s) / (reservationPrice ν s - s)) :
    (∀ p q, 0 ≤ q →
        fullRefundProfit ν D c s p q ≤ profit ν D c s (reservationPrice ν s) qP s) ∧
      (StrictMonoOn (cdf D) (Set.Ici 0) → qF ≤ qP) := by sorry

end SuReturns.PartialRefunds
Source
Su, Consumer Returns Policies and Supply Chain Performance, SSRN 1020451 (version of August 2008), p. 12, Corollary 1; proof p. 26
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 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