Theorem 9, p. 553 — closed form of the entropy-constrained worst-case VaR
ProvedWorstCaseVaR.Entropy.entropy_constrained_varLet be a Gaussian distribution on with positive definite covariance , let , and let be the set of probability distributions of returns with Kullback–Leibler divergence
Let be a nonzero portfolio and a loss probability level. The worst-case Value-at-Risk
exists and equals
where , , and is the standard normal cumulative distribution function.
The theorem shows that an entropy ball around a Gaussian changes the Gaussian VaR only through its risk factor: the worst-case VaR keeps the mean–standard deviation form, with replaced by , so it can be optimised over portfolios by second-order cone programming.
Formalization Note "min" is encoded as IsLeast of the feasible set , and "sup over is at most " as "for every ". The paper's range is narrowed to : at every is feasible and no minimum exists. The hypothesis is the paper's standing assumption that the admissible set of portfolios does not contain . is written , and .
import Mathlib import Definitions.Def_WorstCaseVaR_Entropy_Basic
namespace WorstCaseVaR.Entropy
/-- Theorem 9, p. 553: when the distribution of returns is only known to satisfy the relative
entropy constraint `KL(P, P₀) ≤ d` with respect to the Gaussian `P₀ = 𝒩(x̂, Γ)`, `Γ ≻ 0`,
the entropy-constrained worst-case Value-at-Risk (4) of a portfolio `w ≠ 0` at level
`ε ∈ (0, 1)` is `V_𝒫(w) = κ(ε, d)√(wᵀΓw) - x̂ᵀw` (Eq. (44)): this value is the least `γ`
with `P{γ ≤ -wᵀx} ≤ ε` for every such `P`. -/
theorem entropy_constrained_var {n : ℕ} (xhat w : Returns n) (Γ : Matrix (Fin n) (Fin n) ℝ)
(hΓ : Γ.PosDef) (hw : w ≠ 0) (ε d : ℝ) (hε0 : 0 < ε) (hε1 : ε < 1) (hd : 0 ≤ d) :
IsLeast (varFeasible (klBall xhat Γ d) w ε)
(kappaEntropy ε d * Real.sqrt (quadForm Γ w) - inner ℝ xhat w) := by sorry
end WorstCaseVaR.Entropy
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Throughout, is a natural number and is a type defined in the imported module Definitions.Def_WorstCaseVaR_Entropy_Basic, which is not shown here. From how it is used, it has a zero element and a real inner product . Four other names in the statement come from that same unseen module: klBall, varFeasible, kappaEntropy and quadForm. Their definitions are not available to this read-back. They are described below only by their inputs and the type of their output, and are not expanded. The code carries a documentation comment describing what these objects are meant to be, but that comment is not part of what the theorem asserts, and this read-back does not rely on it.
Data. The theorem takes:
- a natural number (implicit);
- two elements ;
- a real matrix , indexed by ;
- two real numbers and .
Hypotheses.
- is positive definite in Mathlib's sense: it is symmetric (Hermitian over ), and for every nonzero .
- in .
- .
- .
Objects built from the definitions module.
- is built from , and .
- is a set of real numbers built from , and .
- is a real number.
- is a real number.
Conclusion. Let
Here is Mathlib's real square root, which returns for negative arguments. The theorem asserts that is the least element of . That means both of the following hold:
- ;
- for every .
This is a statement of attainment, not just an infimum: must itself belong to . The statement does not say what membership in means. That depends entirely on the unseen definitions of varFeasible and klBall, and likewise the value of depends on the unseen kappaEntropy and quadForm.
Degenerate cases.
- . Here is the empty matrix, and "positive definite" holds vacuously. If has only one element, as a zero-dimensional space would, then cannot be satisfied and the theorem is vacuous for . Whether is a one-element type cannot be confirmed from the code shown.
- . This is allowed. Then is whatever
klBallreturns with radius . - . It is kept strictly inside , so the endpoints and are excluded.
- Negative . If were negative, would silently be and would reduce to . Whether can be negative under these hypotheses depends on how
quadFormis defined. - Junk values inside the definitions. Any default values produced inside
kappaEntropy,klBallorvarFeasiblecannot be identified from this code. Examples are division by zero, a logarithm or inverse of a boundary value, or the supremum or infimum of an empty or unbounded set. - Empty . If is empty for some admissible inputs, the conclusion is false there. So the theorem, if true, implies that is nonempty and contains for every admissible choice of .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.