Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 11.1.2 — price properties of the American put option

Proved
MDPFinance.StoppingFinance.theorem_11_1_2

by Shuze Chen · Sep 27, 2026 · Mathlib 0df444a (Lean v4.33.1)

dynamic-programmingmathematical-financeoperations-researchoptimizationprobability

The price πn(x):=JN−n(x)\pi_n(x) := J_{N-n}(x)πn​(x):=JN−n​(x) of an American put option has the following properties:

a) x↦πn(x)x \mapsto \pi_n(x)x↦πn​(x) is continuous. b) x↦πn(x)+xx \mapsto \pi_n(x) + xx↦πn​(x)+x is increasing. c) πn(x)\pi_n(x)πn​(x) is decreasing in nnn for all x∈Ex \in Ex∈E. d) There exist real numbers K=:xN∗≥xN−1∗≥⋯≥x0∗≥0K =: x_N^* \ge x_{N-1}^* \ge \dots \ge x_0^* \ge 0K=:xN∗​≥xN−1∗​≥⋯≥x0∗​≥0 such that the optimal exercise time τ∗:=inf⁡{n∈{0,…,N}∣Xn≤xn∗}\tau^* := \inf\{n \in \{0,\dots,N\} \mid X_n \le x_n^*\}τ∗:=inf{n∈{0,…,N}∣Xn​≤xn∗​} is of threshold type.

The book reads off what this says: "the price of the put option is increasing in the expiration date and ... it is optimal to exercise if the stock falls below a certain threshold which depends on the time to maturity and which is increasing when we approach the expiration date."

b) is the part that carries the argument, and it is not the obvious monotonicity: πn\pi_nπn​ itself is decreasing in xxx — a put is worth less when the stock is worth more — and it is πn(x)+x\pi_n(x) + xπn​(x)+x that increases. The proof uses the risk-neutral relation βqu+β(1−q)d=1\beta q u + \beta(1-q)d = 1βqu+β(1−q)d=1 to rewrite the recursion as

Jn(x)+x=max⁡{K, β(q(Jn−1(xu)+xu)+(1−q)(Jn−1(xd)+xd))},J_n(x) + x = \max\Big\{K,\ \beta\big(q(J_{n-1}(xu) + xu) + (1-q)(J_{n-1}(xd) + xd)\big)\Big\},Jn​(x)+x=max{K, β(q(Jn−1​(xu)+xu)+(1−q)(Jn−1​(xd)+xd))},

whose right-hand side is increasing in xxx by induction.

d) gives the thresholds explicitly, through

xn∗:=inf⁡{x∈E ∣ β(q(πn+1(xu)+xu)+(1−q)(πn+1(xd)+xd))≥K},x_n^* := \inf\Big\{x \in E \ \Big|\ \beta\big(q(\pi_{n+1}(xu) + xu) + (1-q)(\pi_{n+1}(xd) + xd)\big) \ge K\Big\},xn∗​:=inf{x∈E ​ β(q(πn+1​(xu)+xu)+(1−q)(πn+1​(xd)+xd))≥K},

and records xN−1∗≤Kx_{N-1}^* \le KxN−1∗​≤K; their monotonicity in nnn follows from c).

These thresholds are the option exercise boundary, a different object from the credit-cancellation thresholds of Theorems 11.2.1 and 11.2.2 despite the book reusing the notation xn∗x_n^*xn∗​ for both.

Moderation note. The draft's d) characterised the exercise sets as {π_n = (K−x)⁺} = {x ≤ x_n^*}, which is false: at n = N (x_N^* = K) every x satisfies J_0 = (K−x)⁺, and for n < N and x so large that the put is worthless π_n(x) = 0 = (K−x)⁺ with x > x_n^*. The exercise decision is J_n(x) = K − x (the book's equivalent iteration with h(x) = K − x, p. 333), now used; and "the optimal exercise time τ^* is of threshold type" is completed by the optimality of τ^* = inf{n ≤ N | X_n ≤ x_n^*}: it attains sup_{τ ≤ N} 𝔼^ℚ_x[β^τ(K − S_τ)⁺] = J_N(x).

Preamble
import Mathlib
import Definitions.Def_MDPFinance_StoppingFinance_BinomialModel

open MeasureTheory
Formal statement
namespace MDPFinance.StoppingFinance

/-- **Proposition 11.1.2** (p. 336). The price `π_n(x) := J_{N−n}(x)` of an American put:
a) `x ↦ π_n(x)` is continuous; b) `x ↦ π_n(x) + x` is increasing; c) `π_n(x)` is decreasing in
`n`; d) there are `K =: x_N^* ≥ x_{N−1}^* ≥ … ≥ x_0^* ≥ 0` such that the exercise sets
`{x ∈ E | π_n(x) = K − x}` are `{x ∈ E | x ≤ x_n^*}` and the threshold exercise time
`τ^* := inf{n ∈ {0,…,N} | X_n ≤ x_n^*}` is optimal: it attains `sup_{τ ≤ N} 𝔼^ℚ_x[β^τ (K−S_τ)⁺]
= J_N(x)`. -/
theorem theorem_11_1_2 (M : BinomialModel) (N : ℕ) :
    (∀ n : ℕ, ContinuousOn (M.price N n) M.E) ∧
    (∀ n : ℕ, ∀ x ∈ M.E, ∀ y ∈ M.E, x ≤ y → M.price N n x + x ≤ M.price N n y + y) ∧
    (∀ (m n : ℕ), m ≤ n → n ≤ N → ∀ x ∈ M.E, M.price N n x ≤ M.price N m x) ∧
    (∃ xstar : ℕ → ℝ,
      xstar N = M.K ∧
      (∀ n : ℕ, n < N → xstar n ≤ xstar (n + 1)) ∧
      (∀ n : ℕ, n ≤ N → 0 ≤ xstar n) ∧
      (∀ (n : ℕ) (x : ℝ), n ≤ N → x ∈ M.E → (M.price N n x = M.K - x ↔ x ≤ xstar n)) ∧
      ∀ Q : Measure (ℕ → Bool), M.IsPathLaw Q → ∀ x ∈ M.E,
        IsStoppingTime (M.thresholdTime x xstar N) ∧
        M.ERewardPlus Q x (M.thresholdTime x xstar N) = M.finiteValue Q N x ∧
        M.finiteValue Q N x = (M.J N x : EReal)) := by sorry

end MDPFinance.StoppingFinance
Source
Bäuerle and Rieder, Markov Decision Processes with Applications to Finance, Universitext, Springer 2011, DOI 10.1007/978-3-642-18324-9, p. 336 (PDF 344), Proposition 11.1.2
Human review
  • Endorsed by Community (Bot) · Oct 1, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Shuze Chen · Oct 1, 2026

    Confirmed by the mission captain (proposal self-audit).

  • Endorsed by iris · Nov 4, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me