Proof of Proposition 3, p. 32 — ∂Π/∂α_L of the benchmark and of (8)
ProvedSellingData.AppendsBinary.deriv_alphaLcomparative-staticsinformation-sellingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1screening
In the data-append model, let be arbitrary and let be a point at which the selling cost has derivative . Then the low-cost benchmark profit of (4) has derivative
at , and the objective of (8), as a function of , has derivative
at .
The two derivatives differ by the constant , which is negative when and positive when ; this comparison is what drives the direction of the distortion in Proposition 3.
Preamble
import Mathlib import Definitions.Def_SellingData_AppendsBinary_Setting
Formal statement
namespace SellingData.AppendsBinary
open Model
/-- Proof of Proposition 3, p. 32: the derivative in `α_L` of the benchmark profit (4) of the
low-cost type is `(1 − θ)[φ(ū − c_L) − (1 − φ)(u̲ − c_L) − k′(α_L)]`, and that of the objective (8)
is `(1 − θ)[φ(ū − c_L) − (1 − φ)(u̲ − c_L) − k′(α_L)] + θ(1 − 2φ)(c_L − c_H)`. -/
theorem deriv_alphaL (M : Model) (αH a k' : ℝ) (ha : a ∈ Set.Icc M.aLo M.aHi)
(hk : HasDerivAt M.k k' a) :
HasDerivAt (fun α => M.benchProfit SellingData.ListsBinary.CostType.L α)
((1 - M.θ) * (M.φ * (M.uHi - M.cL) - (1 - M.φ) * (M.uLo - M.cL) - k')) a ∧
HasDerivAt (fun α => M.reducedProfit αH α)
((1 - M.θ) * (M.φ * (M.uHi - M.cL) - (1 - M.φ) * (M.uLo - M.cL) - k')
+ M.θ * (1 - 2 * M.φ) * (M.cL - M.cH)) a := by sorry
end SellingData.AppendsBinary
Source
Guo, Selling Data to Marketers, author's manuscript (Management Science, doi:10.1287/mnsc.2023.01350), p. 32, proof of Proposition 3, display for ∂Π/∂α_L
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.