(A.27)–(A.28), p. 21 — strict log-supermodularity identity for Firm 1
ProvedAccelScal.UCRace.eq_A28log-supermodularityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1r-and-dstrategic-complements
For positive investment intensities and discount rate , let be Firm 1's closed-form value from (9). The cross-partial expression used to test log-supermodularity satisfies
The expression is strictly positive when both firms' discovery rates rise with their investment intensities and Firm 1's project surplus is nonzero. This is the paper's strict log-supermodularity calculation for the race.
Formalization Note Under the power-form technology, the derivative conditions correspond to . The nonzero-surplus hypothesis is necessary for strictness: the squared factor vanishes at . The identity itself holds without that extra condition.
Preamble
import Mathlib import Definitions.Def_AccelScal_UCRace_Setting noncomputable section
Formal statement
namespace AccelScal.UCRace
/-- Lin, Appendix A.6, equations (A.27)–(A.28), p. 21. The strict
inequality requires a nonzero project surplus; it fails at the upper endpoint. -/
theorem eq_A28 (T1 T2 : Tech) (h1 : T1.Standing) (h2 : T2.Standing)
(r : ℝ) (hr : 0 < r) (R1 R2 : ℝ) (hR1 : 0 < R1) (hR2 : 0 < R2) :
(let V := V1 T1 T2 r
let lhs := V R1 R2 * deriv (fun b => deriv (fun a => V a b) R1) R2
- deriv (fun a => V a R2) R1 * deriv (fun b => V R1 b) R2
lhs = deriv T2.lam R2 * (T1.u R1 * T1.lam R1 - R1) ^ 2 /
(r + T1.lam R1 + T2.lam R2) ^ 4 * deriv T1.lam R1 ∧
(0 < T1.γ → 0 < T2.γ → T1.u R1 * T1.lam R1 ≠ R1 → 0 < lhs)) := by sorry
end AccelScal.UCRace
Source
Lin, Accelerability vs. Scalability: R&D Investment Under Financial Constraints and Competition, Management Science, Articles in Advance (2022), doi:10.1287/mnsc.2022.4503, Appendix A.6, p. 21, (A.27)–(A.28)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.