Deterministic linear drift subtraction preserves the actual spectrally negative Lévy law and shifts its exponent
ProvedAvramDividend.Classical.levy_linear_drift_shift_exists_ws01favram-classicaldrift-shiftlevy-processrogozin-regularitywiener-hopf-interfacews01f
For every original spectrally negative Lévy process and arbitrary real slope C, construct an actual spectrally negative Lévy process on the same filtered probability space whose sample paths are X_t−Ct, Gaussian and Lévy-measure parameters are unchanged, triplet drift coefficient decreases by C, and Laplace exponent becomes ψ(θ)−Cθ. The proof must reverify adaptedness, stationary independent increments, càdlàg paths, no positive jumps and the source exponential moment identity. This is the independent, source-faithful stochastic interface needed to deduce all cofinal Rogozin slopes from an eventual descending Wiener–Hopf factor.
Preamble
import Mathlib import Definitions.Def_AvramDividend_Classical_SpectrallyNegativeLevy open MeasureTheory ProbabilityTheory Filter Set Topology open scoped NNReal ENNReal open AvramDividend.Classical
Formal statement
theorem AvramDividend.Classical.levy_linear_drift_shift_exists_ws01f
{Ω : Type*} [mΩ : MeasurableSpace Ω]
{P : Measure Ω} {𝓕 : Filtration ℝ≥0 mΩ}
(X : SpectrallyNegativeLevy P 𝓕) (C : ℝ) :
∃ Y : SpectrallyNegativeLevy P 𝓕,
(∀ (t : ℝ≥0) (ω : Ω),
Y.X t ω = X.X t ω - C * (t : ℝ)) ∧
(∀ θ : ℝ, Y.ψ θ = X.ψ θ - C * θ) := by sorrySource
Canonical definition UUID 933ced80-71c1-4c0f-a6a7-ab7757fc97e9 and elementary deterministic affine transformation of a Lévy process; Mathlib IdentDistrib.sub_const and independence of measurable transforms.