Theorem 2 (Chatterjee–Samuelson): class equilibrium strategies satisfy the linked differential equations (3a)–(3b)
ProvedChatterjeeSamuelson.LinkedODE.linked_differential_equationsConsider sealed-offer bargaining between a seller with reservation price and a buyer with reservation price : trade occurs iff the buyer's offer is at least the seller's ask , at price with . The buyer's belief about is a regular belief with distribution function and density ; the seller's belief about is a regular belief with distribution function and density . Let be an equilibrium in which both strategies are of class . Then:
- (3a) For every seller value in an open interval on which is strictly increasing, and every buyer value with ,
- (3b) For every buyer value in an open interval on which is strictly increasing, and every seller value with ,
Here in (3a) is the paper's and in (3b) is its . The theorem turns the equilibrium problem into a pair of ordinary differential equations that link the two strategies; the paper's explicit equilibria (Example 1 for uniform beliefs) are solutions of this system.
Formalization Note is Lean's deriv S y, which the class assumption makes a genuine derivative at such ; no sign condition on is assumed. The inverses , are not introduced as functions: the matching value is a bound variable, so the equations are asserted exactly where the inverse is defined.
import Mathlib import Definitions.Def_ChatterjeeSamuelson_LinkedODE_RegularBelief import Definitions.Def_ChatterjeeSamuelson_LinkedODE_ClassA import Definitions.Def_ChatterjeeSamuelson_Shared_IsEquilibrium open MeasureTheory ProbabilityTheory Set
namespace ChatterjeeSamuelson.LinkedODE
/-- **Theorem 2** (Chatterjee & Samuelson, *Bargaining under Incomplete Information*,
Oper. Res. 31(5) 1983, §2, p. 840 [PDF 6]): "In a class A equilibrium, over intervals of
reservation prices for which the offer strategies are strictly increasing, S( ) and B( )
must satisfy the linked differential equations
kF_b(y)S′(y) + f_b(y)S(y) = B⁻¹(S(y))f_b(y) (3a)
(1 − k)(1 − F_s(x))B′(x) − f_s(x)B(x) = −S⁻¹(B(x))f_s(x). (3b)"
Setting: `0 ≤ k ≤ 1`; seller values in `[loS, hiS]`, buyer values in `[loB, hiB]`; the
buyer's belief `μb` about `v_s` is regular on `[loS, hiS]` with CDF `F_b = cdf μb` and
density `fb`; the seller's belief `μs` about `v_b` is regular on `[loB, hiB]` with
`F_s = cdf μs` and density `fs`; `(S, B)` is an equilibrium and both strategies are of
class `A`. Then
* (3a) holds at every seller value `y` interior to an open interval `(a, c) ⊆ [loS, hiS]`
on which `S` is strictly increasing, for every buyer value `x ∈ [loB, hiB]` with
`B x = S y` (the paper's `B⁻¹(S(y))`);
* (3b) holds at every buyer value `x` interior to an open interval `(a, c) ⊆ [loB, hiB]`
on which `B` is strictly increasing, for every seller value `y ∈ [loS, hiS]` with
`S y = B x` (the paper's `S⁻¹(B(x))`).
*Formalization Note.* `S′(y)` is `deriv S y`; class `A` makes `S` differentiable there.
No sign condition on `S′(y)` is assumed. `B⁻¹` and `S⁻¹` are not introduced as functions:
the inverse image is a bound variable, so the equations are asserted exactly where the
paper's inverse is defined. -/
theorem linked_differential_equations (k : ℝ) (hk0 : 0 ≤ k) (hk1 : k ≤ 1)
(μb μs : Measure ℝ) (loS hiS loB hiB : ℝ) (fb fs S B : ℝ → ℝ)
(hμb : RegularBelief μb loS hiS fb) (hμs : RegularBelief μs loB hiB fs)
(hEq : Shared.IsEquilibrium k μb μs loS hiS loB hiB S B)
(hS : ClassA S loS hiS) (hB : ClassA B loB hiB) :
(∀ a c y : ℝ, loS ≤ a → c ≤ hiS → y ∈ Ioo a c → StrictMonoOn S (Ioo a c) →
∀ x ∈ Icc loB hiB, B x = S y →
k * cdf μb y * deriv S y + fb y * S y = x * fb y) ∧
(∀ a c x : ℝ, loB ≤ a → c ≤ hiB → x ∈ Ioo a c → StrictMonoOn B (Ioo a c) →
∀ y ∈ Icc loS hiS, S y = B x →
(1 - k) * (1 - cdf μs x) * deriv B x - fs x * B x = -(y * fs x)) := by sorry
end ChatterjeeSamuelson.LinkedODE
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.