Proposition 5(i), Online App. p. 8 — for fixed protection C, a purely usage-driven business's data collection is socially optimal
ProvedDigitalPrivacy.Welfare.proposition_5_iWork in the model of DigitalPrivacy.Welfare.Setting with a purely usage-driven business, , so that the profit is . Assume the printed ranges , , and (the case is allowed). Fix a protection level and a collection level . Then maximizes the business's profit over data collection at protection if and only if it maximizes the planner's objective at protection :
In words: for a given level of data protection, the data collection chosen by a purely usage-driven business is socially optimal. In the proof of Theorem 2 this is what lets the comparison of the two optima for run through a single best-response map in .
Formalization Note. The paper states Proposition 5 for a general profit function (Online App. C.5); for the linear profit (2), "purely usage-driven" is . The paper's proof concludes , i.e. the two sets of maximizers in coincide, which is the equivalence stated here. Only the first claim of Proposition 5(i) (its first two sentences) is formalized.
import Mathlib import Definitions.Def_DigitalPrivacy_Welfare_Setting
namespace DigitalPrivacy.Welfare
theorem proposition_5_i (β ρ ψ α : ℝ) (h : PrintedBounds β ρ ψ 0 α) (C : ℝ) (hC : 0 ≤ C)
(ξ : ℝ) (hξ : ξ ∈ Set.Icc (0 : ℝ) 1) :
(∀ ξ' ∈ Set.Icc (0 : ℝ) 1, Pi β ρ ψ 0 ξ' C ≤ Pi β ρ ψ 0 ξ C) ↔
(∀ ξ' ∈ Set.Icc (0 : ℝ) 1, W β ρ ψ 0 α ξ' C ≤ W β ρ ψ 0 α ξ C) := by sorry
end DigitalPrivacy.Welfare
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.