Stochastic representation and path properties of a fractional Cox–Ingersoll–Ross process

We consider the Cox–Ingersoll–Ross process that satisfies the stochastic differential equation dXt=aXtdt+σXtdBtHdX_t = aX_t dt+\sigma \sqrt {X_t} dB^H_t driven by a fractional Brownian motion BtHB^H_t with the Hurst index exceeding 23\frac {2}{3}, where ∫0tXsdBsH\int _0^t\sqrt {X_s} dB^H_s is the pa...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Theory of probability and mathematical statistics 2019-02, Vol.97, p.167-182
Hauptverfasser: Mishura, Yu. S., Piterbarg, V. I., Ralchenko, K. V., Yurchenko-Tytarenko, A. Yu
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We consider the Cox–Ingersoll–Ross process that satisfies the stochastic differential equation dXt=aXtdt+σXtdBtHdX_t = aX_t dt+\sigma \sqrt {X_t} dB^H_t driven by a fractional Brownian motion BtHB^H_t with the Hurst index exceeding 23\frac {2}{3}, where ∫0tXsdBsH\int _0^t\sqrt {X_s} dB^H_s is the pathwise integral defined as the limit of the corresponding Riemann–Stieltjes sums. We show that the Cox–Ingersoll–Ross process coincides with the square of the fractional Ornstein–Uhlenbeck process up to the first return to zero. Based on this observation, we consider the square of the fractional Ornstein–Uhlenbeck process with an arbitrary Hurst index and prove that it satisfies the above stochastic differential equation up to the first return to zero if ∫0tXsdBsH\int _0^t\sqrt {X_s} dB^H_s is understood as the pathwise Stratonovich integral. Then a natural question arises about the first visit to zero of the fractional Cox–Ingersoll–Ross process which coincides with the first visit to zero of the fractional Ornstein–Uhlenbeck process. Since the latter process is Gaussian, we use the bounds for the distributions of Gaussian processes to prove that the probability of a visit to zero over a finite time equals 1 if a>0a>0. Otherwise this probability is positive. We provide an upper bound for this probability.
ISSN:0094-9000
1547-7363
DOI:10.1090/tpms/1055