Proposition 2 — the simultaneous KS equilibrium is unique, equals resp. , and is the NB equilibrium exactly at retailer power
ProvedCompetingKS.proposition_2Let and .
One-to-two channel. Let and .
- The simultaneous Kalai–Smorodinsky equilibrium exists, is unique, and equals .
- For , is a simultaneous Nash bargaining equilibrium with retailer bargaining power if and only if .
- The simultaneous Nash bargaining equilibrium with exists, is unique, and equals .
Two-to-one channel. Let and . The same three statements hold with in place of .
So in both channels the Kalai–Smorodinsky equilibrium coincides with the Nash bargaining equilibrium in which the retailer has power , not the symmetric , whatever the level of competition . This is the paper's central finding under simultaneous negotiation without contingency.
Formalization Note A simultaneous equilibrium is a contract pair in (resp. ) in which each contract is the solution of its bargaining unit given the other contract, over contracts in . The Kalai–Smorodinsky solution is individual rationality, Pareto optimality and the proportional-surplus equation.
Differences from the source:
- encoding: η ∈ [0, 1] for the two-to-one channel → η ∈ [0, 1); reason: at η = 1, wᵢ = wⱼ < a Lemma 1(ii)(a) gives qᵢ = 0 and (ii)(c) gives qᵢ = (a − wᵢ)/(2b), and (ii)(b) divides by 1 − η² = 0, so q* and the profits are undefined there; this excludes only that undefined case
- reading: the negotiated wholesale prices are ŵ, w̌ → the simultaneous KS equilibrium exists, is unique and equals (ŵ, ŵ), resp. (w̌, w̌); reason: "the negotiated prices are" characterizes the equilibrium; the stronger reading; equilibrium is "the Nash equilibrium of the two bargaining problems", p.19
- reading: prices "equal to the corresponding NB solutions with the retailers' bargaining power being 0.6" → (ŵ, ŵ) is an NB(θ) equilibrium for θ ∈ [0, 1] if and only if θ = 3/5, and the NB(3/5) equilibrium is exactly that pair; same for (w̌, w̌); reason: the proof sets the NB equilibrium price equal to the KS price and solves for θ
- reading: the KS solution as equation (2) on a non-convex allocation set → individual rationality, Pareto optimality and (2) cross-multiplied; reason: the paper restricts (2) to Pareto allocations (online appendix p.5, pp.14–15); without Pareto every contract with allocation (D, d) satisfies (2) as 0 = 0
- encoding: contract space, in the channels → w ∈ [c, a]; reason: the contract space stated on p.15
- encoding: case conditions as ratios, with 2/η and 1/η → the same conditions with cleared denominators; reason: the ratios divide by η and by a − wⱼ; the cleared forms agree with the printed ones wherever those are defined and read 2/η as +∞ at η = 0, which changes no instance
- generalization: equilibria range over Ĉ, Č of (6), defined for wⱼ < a → the cleared forms also admit pairs with a coordinate equal to a; reason: uniqueness is then claimed over a larger set, which contains the source's; no such pair is an equilibrium
import Mathlib import Definitions.Def_CompetingKS_unitOT
namespace CompetingKS
/-- Proposition 2 (Simultaneous Bargaining without Contingency), Feng, Li, Shanthikumar,
*Negotiations in Competing Supply Chains: The Kalai-Smorodinsky Bargaining Solution*, accepted
manuscript (Management Science, July 3, 2021), SSRN 3447674, pp.19–20.
For every `a > c`, `b > 0`. In the one-to-two channel, for every `η ∈ [0, 1]`, with
`ŵ = c + (a - c)/5`: the simultaneous KS equilibrium exists, is unique and equals `(ŵ, ŵ)`; `(ŵ, ŵ)`
is a simultaneous NB equilibrium with retailer power `θ ∈ [0, 1]` exactly when `θ = 3/5`; and the NB
equilibrium at `θ = 3/5` exists, is unique and equals `(ŵ, ŵ)`. In the two-to-one channel, for every
`η ∈ [0, 1)`, with `w̌ = c + (1 - η)(a - c)/(5 - η)`: the same three statements.
**Formalization Note** The two-to-one channel excludes `η = 1`, where the claimed `w̌ = c` lies on
the diagonal on which Lemma 1(ii) does not define `q*`. Equilibria are `IsKSEqOT`, `IsNBEqOT`,
`IsKSEqTO`, `IsNBEqTO`; the KS solution is `IsKSResponse`, individual rationality, Pareto optimality
and equation (2). -/
theorem proposition_2 {a b c : ℝ} (hac : c < a) (hb : 0 < b) :
(∀ η : ℝ, 0 ≤ η → η ≤ 1 →
(∀ w₁ w₂ : ℝ, IsKSEqOT a b c η w₁ w₂ ↔ w₁ = c + (a - c) / 5 ∧ w₂ = c + (a - c) / 5) ∧
(∀ θ : ℝ, 0 ≤ θ → θ ≤ 1 →
(IsNBEqOT a b c η θ (c + (a - c) / 5) (c + (a - c) / 5) ↔ θ = 3 / 5)) ∧
(∀ w₁ w₂ : ℝ, IsNBEqOT a b c η (3 / 5) w₁ w₂ ↔
w₁ = c + (a - c) / 5 ∧ w₂ = c + (a - c) / 5)) ∧
(∀ η : ℝ, 0 ≤ η → η < 1 →
(∀ w₁ w₂ : ℝ, IsKSEqTO a b c η w₁ w₂ ↔
w₁ = c + (1 - η) * (a - c) / (5 - η) ∧ w₂ = c + (1 - η) * (a - c) / (5 - η)) ∧
(∀ θ : ℝ, 0 ≤ θ → θ ≤ 1 →
(IsNBEqTO a b c η θ (c + (1 - η) * (a - c) / (5 - η))
(c + (1 - η) * (a - c) / (5 - η)) ↔ θ = 3 / 5)) ∧
(∀ w₁ w₂ : ℝ, IsNBEqTO a b c η (3 / 5) w₁ w₂ ↔
w₁ = c + (1 - η) * (a - c) / (5 - η) ∧ w₂ = c + (1 - η) * (a - c) / (5 - η))) := by sorry
end CompetingKS
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.