App. A.4, p. 34 — for P = 0, a business that over-collects also over-protects
ProvedDigitalPrivacy.Welfare.usage_driven_overcollectioncomparative-staticsdata-privacyp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1social-welfare
Work in the model of DigitalPrivacy.Welfare.Setting with a purely usage-driven business, , under the standing assumptions , , , . Let be any business optimum and any social optimum. If the business over-collects relative to the planner, it also over-protects:
This is the first step of the purely usage-driven part of the proof of Theorem 2 (App. A.4), where the paper derives it "from the first-order conditions on the data collection". Combined with the argument behind (11), it rules out over-collection for .
Formalization Note. The standing assumptions add and to the paper's printed ranges, as in Theorem 2.
Preamble
import Mathlib import Definitions.Def_DigitalPrivacy_Welfare_Setting
Formal statement
namespace DigitalPrivacy.Welfare
theorem usage_driven_overcollection (β ρ ψ α : ℝ) (h : Standing β ρ ψ 0 α) (ξs Cs ξW CW : ℝ)
(hs : DigitalPrivacy.Strategy.IsBusinessOpt β ρ ψ 0 ξs Cs) (hW : IsSocialOpt β ρ ψ 0 α ξW CW)
(hover : ξW < ξs) :
CW < Cs := by sorry
end DigitalPrivacy.Welfare
Source
Fainmesser, Galeotti, Momot, Digital Privacy, author manuscript (This Draft: April 2022; SSRN 3459274), p. 34, App. A.4, last paragraph (proof of Theorem 2, case P = 0)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.