English

Invariance times

Computational Finance 2017-02-06 v1 Pricing of Securities

Abstract

On a probability space (Ω,A,Q)(\Omega,\mathcal{A},\mathbb{Q}) we consider two filtrations FG\mathbb{F}\subset \mathbb{G} and a G\mathbb{G} stopping time θ\theta such that the G\mathbb{G} predictable processes coincide with F\mathbb{F} predictable processes on (0,θ](0,\theta]. In this setup it is well-known that, for any F\mathbb{F} semimartingale XX, the process XθX^{\theta-} (XX stopped "right before θ\theta") is a G\mathbb{G} semimartingale.Given a positive constant TT, we call θ\theta an invariance time if there exists a probability measure P\mathbb{P} equivalent to Q\mathbb{Q} on F_T\mathcal{F}\_T such that, for any (F,P)(\mathbb{F},\mathbb{P}) local martingale XX, XθX^{\theta-} is a (G,Q)(\mathbb{G},\mathbb{Q}) local martingale. We characterize invariance times in terms of the (F,Q)(\mathbb{F},\mathbb{Q}) Az\'ema supermartingale of θ\theta and we derive a mild and tractable invariance time sufficiency condition. We discuss invariance times in mathematical finance and BSDE applications.

Cite

@article{arxiv.1702.01045,
  title  = {Invariance times},
  author = {Stéphane Crépey and Shiqi Song},
  journal= {arXiv preprint arXiv:1702.01045},
  year   = {2017}
}
R2 v1 2026-06-22T18:08:43.543Z