Discounted surviving first-record payments dominate the left scale-ratio geometric sequence
OpenAvramDividend.Classical.killed_firstRecord_grid_expected_scale_lowerFor a standing spectrally negative Lévy process and its positive q-scale function W, let ε be a record-height increment smaller than the ruin barrier a, and suppose 0 ≤ W(a−ε) < W(a). For every n, the expected discounted payment of size ε at the (n+1)st strict upward record crossing, killed if ruin precedes that crossing, is at least ε times the (n+1)st power of W(a−ε)/W(a). This is the stochastic first-passage and strong-Markov renewal estimate missing from the independently Proved pathwise record-grid budget.
Preamble
import Mathlib.MeasureTheory.Integral.Lebesgue.Basic import Mathlib.Data.ENNReal.Real import Mathlib.Analysis.Complex.Exponential import Definitions.Def_AvramDividend_Classical_SpectrallyNegativeLevy import Definitions.Def_AvramDividend_Classical_ScaleFunction import Definitions.Def_AvramDividend_Classical_ReflectionBarrier import Theorems.Thm_AvramDividend_Classical_chat08_record_grid_stopped_weights_aemeasurable open MeasureTheory open scoped NNReal ENNReal open AvramDividend.Classical
Formal statement
theorem AvramDividend.Classical.killed_firstRecord_grid_expected_scale_lower
{Ω : Type*} [mΩ : MeasurableSpace Ω] {P : Measure Ω}
{𝓕 : Filtration ℝ≥0 mΩ}
(X : SpectrallyNegativeLevy P 𝓕) (hX : X.Standing)
(q : ℝ) (hq : 0 < q) (W : ℝ → ℝ)
(hW : IsScaleFunction X q W) (a ε : ℝ)
(ha : 0 < a) (hε : 0 < ε) (hεa : ε < a)
(hWa : 0 < W a) (hWleft : 0 ≤ W (a - ε))
(hWinc : W (a - ε) < W a) :
∀ n : ℕ,
ENNReal.ofReal (ε * (W (a - ε) / W a) ^ (n + 1)) ≤
(∫⁻ ω,
if twoSidedExitUpTime X (((n + 1 : ℕ) : ℝ) * ε) ω <
barrierRuinTime X a ω then
ENNReal.ofReal (ε * Real.exp
(-(q * (twoSidedExitUpTime X (((n + 1 : ℕ) : ℝ) * ε) ω).toReal)))
else 0 ∂P) := by sorrySource
Avram, Palmowski and Pistorius (2007), Proposition 1 and the two-sided q-scale exit identity; the missing input is the stopped first-passage Markov regeneration estimate, with the exact indicator and discounting under P.