betaFeedback

WorksMathematics

Berry Esseen bounds for fractional Vasicek type processes: continuous and high-frequency regimes

Abstract

We study the estimation of the drift parameters (a,b)(a,b) of the fractional Vasicek process dXt=a(b−Xt) dt+dBtHdX_t=a(b-X_t)\,dt+dB^H_t, driven by a fractional Brownian motion of Hurst index H∈(0,1)H\in(0,1). The estimators are of moment type, built from the empirical mean and variance of the path, and are analysed in two regimes: continuous observation on [0,T][0,T], and high-frequency observation at times ti=iΔnt_i=iΔ_n with Δn→0Δ_n\to0 and Tn=nΔn→∞T_n=nΔ_n\to\infty. Using the Malliavin–Stein method on the second Wiener chaos, we prove consistency of both estimators for every HH, and quantitative central limit theorems in the Wasserstein distance: for H<34H<\frac34 for the estimator of aa, with explicit rates depending on HH, and for every H∈(0,1)H\in(0,1) for the estimator of bb, at the rate T−HT^{-H}. In the discrete regime the bounds contain an additional discretization term of order Δnmin⁡{H+1/2, 1}Δ_n^{\min\{H+1/2,\,1\}}, and for H<12H<\frac12 the central limit theorem for the estimator of bb requires in addition Tn1−2HΔn1+2H→0T_n^{1-2H}Δ_n^{1+2H}\to0. A short Monte Carlo study illustrates the results.

Introduction

Mean-reverting diffusions are the basic building block of many models in mathematical finance. The Vasicek model of the short interest rate [25] is the Ornstein–Uhlenbeck equation dXt=a(b−Xt) dt+dWtdX_t = a(b-X_t)\,dt + dW_t, in which a>0a > 0 is the speed of mean reversion and bb the long-run level. Replacing the Brownian motion WW by a fractional Brownian motion (fBm) BHB^H of Hurst index H∈(0,1)H \in(0,1) [17] allows the noise to have memory: long memory when H>12H > \frac{1}{2}, as in the continuous-time long-memory volatility models of Comte and Renault [8], and rough, anti-persistent trajectories when H<12H < \frac{1}{2}, which is the regime reported for the log-volatility of financial assets by Gatheral et al. [13]. In this paper we consider the fractional Vasicek process

X0=0,dXt=a(b−Xt) dt+dBtH,t≥0,(1)X_0 = 0,\qquad dX_t = a(b-X_t)\,dt + dB_t^H,\qquad t \ge0, \tag*{(1)}

where a>0a > 0 and b∈Rb \in\mathbb{R} are unknown, H∈(0,1)H \in(0,1) is known, and XX is observed either continuously on [0,T][0,T] or at the discrete times ti=iΔnt_i=i\Delta_n, i=0,…,ni=0,\ldots,n, with Δn→0\Delta_n \to0 and Tn:=nΔn→∞T_n := n\Delta_n \to\infty (high-frequency regime). The process XX is asymptotically stationary and ergodic [7], and bb is its asymptotic mean.

Drift estimation for equations driven by fBm has been studied extensively. For the fractional Ornstein–Uhlenbeck (fOU) process, Hu and Nualart [14] studied a least-squares estimator for H≥12H \ge\frac{1}{2}, proved its strong consistency and a central limit theorem for H∈[12,34)H \in[\frac{1}{2},\frac{3}{4}), and introduced an alternative, simulation-friendly estimator based on the ergodic theorem. Hu et al. [15] extended the study to the whole range H∈(0,1)H \in(0,1), for continuous as well as discrete observations, and treated the case where the volatility is also unknown. Closer to the present setting, Mishura and Ralchenko [18] studied drift estimation in models driven by fBm from discrete observations, and Kukush et al. [16] proved the strong consistency of the drift estimators of the fOU process, for continuous and discrete observations and all H∈(0,1)H \in(0,1). The estimators of the second type, which are also the ones used in this paper, are moment-type estimators based on the ergodic theorem, whose rates are governed by the covariance function of the stationary solution [12]. For the Vasicek-type model (1) with a continuous record, the least-squares estimators were analysed by Xiao and Yu [26] for H≥12H \ge\frac{1}{2} and by Xiao and Yu [27] for H<12H < \frac{1}{2}, and the maximum likelihood estimator by Tanaka et al. [24]; Nourdin and Tran [22] considered Hermite-driven Vasicek-type models and Alazemi et al. [1] studied discrete observations with a fixed step.

The Malliavin–Stein method of Nourdin and Peccati [20], together with the optimal fourth moment theorem [21], has made it possible to obtain explicit rates of convergence in the central limit theorems for estimators of this type. Such Berry–Esséen bounds in the Wasserstein distance were obtained for second-moment estimators of stationary Gaussian processes and for the fOU process by Es-Sebaiy and Viens [12] and Douissi et al. [10], and in the least-squares setting by Chen et al. [5] and, for H<12H < \frac{1}{2}, by Chen and Li [6]. For observations at high frequency, Douissi et al. [9] combined the Berry–Esséen bound of the continuous-time estimator with an estimate of the discretization error. This discretization error has recently been sharpened by Alazemi et al. [2] and Es-Sebaiy and Chen [11], who also treat the Kolmogorov and total variation distances. To our knowledge, however, the quantitative results available in the high-frequency regime concern estimators of a single drift parameter of a mean-zero model, whereas the Vasicek-type model (1) requires the joint estimation of the speed aa and of the level bb.

We construct moment-type estimators (a^,b^)(\hat{a},\hat{b}) of (a,b)(a,b) from the empirical mean and the empirical variance of the path: b^\hat{b} is the time average of XX, which is a first-chaos functional of BHB^H, and a^\hat{a} is obtained by inverting the relation fH(a)=lim⁡Qˉf_H(a)=\lim\bar{Q} (almost surely), where Qˉ\bar{Q} is the empirical variance and fH(a)=HΓ(2H)a−2Hf_H(a)=H\Gamma(2H)a^{-2H}. This lives in the second Wiener chaos up to a negligible term, and our main results are the following.

  • Continuous observation. Both estimators are strongly consistent for every H∈(0,1)H \in(0,1). For H<34H < \frac{3}{4}, we obtain explicit Berry–Esséen bounds in the Wasserstein distance for the empirical variance and for a^\hat{a} (Theorem 3.7), and for every H∈(0,1)H \in(0,1) we show that T1−H(b^T−b)T^{1-H}(\hat{b}_T-b) is at Wasserstein distance O(T−H)O(T^{-H}) of a centered Gaussian law (Theorem 4.2).

  • High-frequency observation. We prove consistency (Theorem 5.5) and Berry–Esséen bounds (Theorem 5.7) in which the continuous-time rate is perturbed by a discretization term of order Δnmin⁡{H+1/2,1}\Delta_n^{\min\{H+1/2,1\}}. For H≥12H \ge\frac{1}{2} no relation between Δn\Delta_n and TnT_n is required for the central limit theorems; for H<12H < \frac{1}{2} the central limit theorem for b^n\hat{b}_n needs Tn1−2HΔn1+2H→0T_n^{1-2H}\Delta_n^{1+2H} \to0 (Corollary 5.8). For H≥12H \ge\frac{1}{2} the discretization term is of the same order as the one obtained for the fOU process in Es-Sebaiy and Chen [11].

  • Technical tools. The estimator a^\hat{a} requires uniform negative moments of the empirical variance. We prove them in the appendix with the Mandelbrot–Van Ness representation of the fBm and the Carbery–Wright inequality [4], for the continuous and the discrete cases.

A short Monte Carlo study (§ 5.5) illustrates the Gaussian approximation and the role of the slow rates near H=34H = \frac{3}{4}.

The paper is organized as follows. § 2 recalls the elements of Malliavin calculus on Wiener space that are used throughout. § 3 treats the continuous-time estimators, § 4 the estimator of bb, and § 5 the high-frequency observations, including the numerical illustration. The technical estimates on the covariance of the stationary solution and on the negative moments of the empirical variance are gathered in the Appendix.

Preliminaries

In this section we recall the elements of Malliavin calculus on Wiener space that are used throughout the paper; we refer to Nourdin and Peccati [20 Chapter 2] and Nualart [23 Chapter 1] for more details.

Fractional Brownian motion and its Gaussian space

Let BH=(BtH)t∈RB^H = (B_t^H)_{t \in\mathbb{R}} be a two-sided fractional Brownian motion of Hurst index H∈(0,1)H \in(0,1), that is, a centered Gaussian process with B0H=0B_0^H = 0 and covariance

E[BtHBsH]=RH(t,s)=12(∣t∣2H+∣s∣2H−∣t−s∣2H).\mathbb{E}[B_t^H B_s^H] = R_H(t,s) = \frac{1}{2}\left(|t|^{2H} + |s|^{2H} - |t-s|^{2H}\right).

Let H\mathfrak{H} be the Hilbert space defined as the closure of the real-valued step functions on R\mathbb{R} with respect to the inner product ⟨1[0,t],1[0,s]⟩H=RH(t,s)\langle\mathbf{1}_{[0,t]}, \mathbf{1}_{[0,s]}\rangle_{\mathfrak{H}} = R_H(t,s). The map 1[0,t]↦BtH\mathbf{1}_{[0,t]} \mapsto B_t^H extends to a linear isometry h↦BH(h)h \mapsto B^H(h) from H\mathfrak{H} onto the Gaussian space generated by BHB^H, so that BH={BH(h):h∈H}B^H = \{B^H(h) : h \in\mathfrak{H}\} is an isonormal Gaussian process over H\mathfrak{H}:

E[BH(h)BH(g)]=⟨h,g⟩H,h,g∈H.\mathbb{E}[B^H(h)B^H(g)] = \langle h,g\rangle_{\mathfrak{H}}, \qquad h,g \in\mathfrak{H}.

For deterministic kernels in H\mathfrak{H}, BH(h)B^H(h) is the Wiener integral ∫Rh dBH\int_{\mathbb{R}} h\,dB^H. It is convenient to also recall the Mandelbrot–Van Ness representation [17]

BtH=∫RK(t,u) dWu,t∈R,K(t,u)=cH[(t−u)+H−12−(−u)+H−12],(2)B_t^H = \int_{\mathbb{R}} K(t,u)\,dW_u,\qquad t \in\mathbb{R},\qquad K(t,u) = c_H\left[(t-u)_+^{H-\frac{1}{2}} - (-u)_+^{H-\frac{1}{2}}\right], \tag*{(2)}

where WW is a two-sided standard Brownian motion and cH>0c_H > 0 is such that Var⁡B1H=1\operatorname{Var} B_1^H = 1.

Wiener chaos and multiple integrals

For q≥1q \ge1 let H⊗q\mathfrak{H}^{\otimes q} and H⊙q\mathfrak{H}^{\odot q} denote the qq-th tensor product of H\mathfrak{H} and its symmetric subspace. The multiple Wiener–Itô integral Iq:H⊙q→L2(Ω)I_q : \mathfrak{H}^{\odot q} \to L^2(\Omega) is defined by Iq(h⊗q)=Hq(BH(h))I_q(h^{\otimes q}) = H_q(B^H(h)) for ∥h∥H=1\lVert h\rVert_{\mathfrak{H}} = 1, where HqH_q is the qq-th Hermite polynomial; I1(h)=BH(h)I_1(h) = B^H(h), and Iq(f)=Iq(f~)I_q(f) = I_q(\widetilde{f}) for f∈H⊗qf \in\mathfrak{H}^{\otimes q}, with f~\widetilde{f} the symmetrization of ff. The qq-th Wiener chaos Hq\mathcal{H}_q is the closed linear span of {Hq(BH(h)):∥h∥H=1}\{H_q(B^H(h)) : \lVert h\rVert_{\mathfrak{H}} = 1\}, and L2(Ω)=⨁q≥0HqL^2(\Omega) = \bigoplus_{q \ge0} \mathcal{H}_q when the underlying σ\sigma-field is generated by BHB^H. We shall use the following properties.

  • Isometry. For f∈H⊙pf \in\mathfrak{H}^{\odot p} and g∈H⊙qg \in\mathfrak{H}^{\odot q},

E[Ip(f)Iq(g)]={p!⟨f,g⟩H⊗pif p=q,0otherwise.(3)\mathbb{E}[I_p(f)I_q(g)] = \begin{cases} p!\langle f,g\rangle_{\mathfrak{H}^{\otimes p}} & \text{if } p=q,\\ 0 & \text{otherwise}. \end{cases} \tag*{(3)}

In particular, elements of Hq\mathcal{H}_q with q≥1q \ge1 are centered, and chaoses of different orders are orthogonal.

References

  1. [1]Alazemi, F., Douissi, S., Es-Sebaiy, K. (2019). Berry–Esséen bounds for drift parameter estimation of discretely observed fractional Vasicek-type process. Theory of Stochastic Processes 24(40), no. 1, 6–18.
  2. [2]Alazemi, F., Es-Sebaiy, K., Al-Foraih, M. (2026). Improved Berry–Esséen bounds for parameter estimation of a fractional Ornstein–Uhlenbeck process observed at high frequency. Advances in Continuous and Discrete Models 2026, 26.DOI
  3. [4]Carbery, A., Wright, J. (2001). Distributional and L^q norm inequalities for polynomials over convex bodies in ℝⁿ. Math. Res. Lett. 8(3), 233–248.DOI
  4. [5]Chen, Y., Kuang, N., Li, Y. (2020). Berry–Esséen bound for the parameter estimation of fractional Ornstein–Uhlenbeck processes. Stochastics and Dynamics 20(4), 2050023.arxiv.org/abs/1806.01487
  5. [6]Chen, Y., Li, Y. (2021). Berry–Esséen bound for the parameter estimation of fractional Ornstein–Uhlenbeck processes with the Hurst parameter H ∈ (0, 1/2). Comm. Statist. Theory Methods 50(13), 2996–3013.
  6. [7]Cheridito, P., Kawaguchi, H., Maejima, M. (2003). Fractional Ornstein–Uhlenbeck processes. Electron. J. Probab. 8, paper no. 3, 1–14.DOI
  7. [8]Comte, F., Renault, E. (1998). Long memory in continuous-time stochastic volatility models. Math. Finance 8(4), 291–323.
  8. [9]Douissi, S., Es-Sebaiy, K., Kerchev, G., Nourdin, I. (2022). Berry–Esséen bounds of second moment estimators for Gaussian processes observed at high frequency. Electron. J. Stat. 16(1), 636–670.arxiv.org/abs/2102.04810
  9. [10]Douissi, S., Es-Sebaiy, K., Viens, F. (2019). Berry–Esséen bounds for parameter estimation of general Gaussian processes. ALEA Lat. Am. J. Probab. Math. Stat. 16, 633–664.arxiv.org/abs/1706.02420
  10. [11]Es-Sebaiy, K., Chen, Y. (2026). New Berry–Esséen bounds for parameter estimation of Gaussian processes observed at high frequency. Preprint, arXiv:2603.05163.
  11. [12]Es-Sebaiy, K., Viens, F. (2019). Optimal rates for parameter estimation of stationary Gaussian processes. Stochastic Process. Appl. 129(9), 3018–3054.
  12. [13]Gatheiral, J., Jaisson, T., Rosenbaum, M. (2018). Volatility is rough. Quant. Finance 18(6), 933–949.
  13. [14]Hu, Y., Nualart, D. (2010). Parameter estimation for fractional Ornstein–Uhlenbeck processes. Statist. Probab. Lett. 80(11–12), 1030–1038.arxiv.org/abs/0901.4925
  14. [15]Hu, Y., Nualart, D., Zhou, H. (2019). Parameter estimation for fractional Ornstein–Uhlenbeck processes of general Hurst parameter. Stat. Inference Stoch. Process. 22(1), 111–142.
  15. [16]Kukush, A., Mishura, Y., Ralchenko, K. (2017). Hypothesis testing of the drift parameter sign for fractional Ornstein–Uhlenbeck process. Electron. J. Stat. 11(1), 385–400.arxiv.org/abs/1604.02645
  16. [17]Mandelbrot, B.B., Van Ness, J.W. (1968). Fractional Brownian motions, fractional noises and applications. SIAM Rev. 10(4), 422–437.DOI
  17. [18]Mishura, Y., Ralchenko, K. (2014). On drift parameter estimation in models with fractional Brownian motion by discrete observations. Austrian J. Statist. 43(3), 218–228.DOI
  18. [20]Nourdin, I., Peccati, G. (2012). Normal Approximations with Malliavin Calculus: From Stein’s Method to Universality. Cambridge Tracts in Mathematics, vol. 192. Cambridge University Press, Cambridge.DOI
  19. [21]Nourdin, I., Peccati, G. (2015). The optimal fourth moment theorem. Proc. Amer. Math. Soc. 143(7), 3123–3133.arxiv.org/abs/1305.1527
  20. [22]Nourdin, I., Tran, T.D. (2019). Statistical inference for Vasicek-type model driven by Hermite processes. Stochastic Process. Appl. 129(10), 3774–3791.arxiv.org/abs/1712.05915
  21. [23]Nualart, D. (2006). The Malliavin Calculus and Related Topics, 2nd edn. Springer, Berlin.
  22. [24]Tanaka, K., Xiao, W., Yu, J. (2020). Maximum likelihood estimation for the fractional Vasicek model. Econometrics 8(3), 32.
  23. [25]Vasicek, O. (1977). An equilibrium characterization of the term structure. J. Financial Economics 5(2), 177–188.
  24. [26]Xiao, W., Yu, J. (2019a). Asymptotic theory for estimating drift parameters in the fractional Vasicek model. Econometric Theory 35(1), 198–231.
  25. [27]Xiao, W., Yu, J. (2019b). Asymptotic theory for rough fractional Vasicek models. Economics Letters 177, 26–29.

Paper details

vlnc:W_nz2eq8qn
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC0-1.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents