betaFeedback

WorksMathematics

Bayesian Inference for Multivariate Stable Distributions via LePage Representation with von Mises – Fisher Spectral Measure

Abstract

We develop Bayesian inference for multivariate αα-stable distributions with von Mises–Fisher spectral measure, motivated by the hitting distribution of Brownian motion on the unit sphere. Inference proceeds via a LePage-type series representation with latent radial, spectral, and Poisson components; truncation bias is absorbed through a Gaussian error term, yielding a likelihood that is smooth in α∈(0,2)α\in(0,2) with no special treatment at the Cauchy boundary α=1α=1. The stability index αα, spectral concentration κκ, mean direction ηη, and spectral mass Γ∞Γ_\infty are estimated jointly by Markov chain Monte Carlo. At the Gaussian boundary α→2α\to 2, the limiting covariance matrix equals twice the second moment matrix of the spectral measure, with explicit parametric form determined by ηη and κκ; in particular, the anisotropy of the limiting Gaussian distribution is controlled solely by the spectral concentration κκ. Simulation across α∈[0.25,1.99]α\in[0.25, 1.99] confirms reliable recovery of all parameters.

Introduction

Stable distributions are increasingly being used in both theoretical and empirical investigations in applied literature. Starting with a well cited article by Mandelbrot [ [21], there has been empirical support for thick tail laws in general and stable laws in particular. McFarland and Sung [1982] used symmetric stable laws to study returns from foreign exchange market; as a continuation of this paper, So [ [33] observed that return distribution follow negatively skewed stable laws. Buckle [ [6] in his analysis of return of a share price observed negatively skewed stable laws. Tsionas [ [34] used sub-Gaussian univariate stable laws in a stable regression setup. Carr and Wu [ [8] used negatively skewed stable laws in analyzing persistence of volatility smirk across maturity spectrum in the context of S&P 500 Index options; they observed that a skewed univariate stable law does a better job of fitting option prices as compared to stochastic volatility models. Rachev and Mittnik [ [29] has extensively used thick tails distributions. Thus, skewed univariate stable law has been receiving the attention of researchers in empirical finance.

There has been increased attention towards multivariate stable distributions in the last two decades. Here again, majority of the applications that we are aware of are from applied finance, mainly with the objective of modeling financial returns from a portfolio. Commonly studied examples are from foreign exchange markets and stock markets, such as, joint movement of a basket of currencies or a set of stock indices. Also, there are occasions where distribution of the basket as a whole is necessary; computation of Value at Risk (VaR) of a portfolio is one such example. Most computations of VaR are currently based on multivariate Gaussian laws, which is known to underestimate losses. Limitations of multivariate Gaussian laws led researchers to practical alternatives, such as historical simulation methods [14, 5], and copula functions [5, 9].

An alternative strand of literature approaches this problem using sub-Gaussian models; see, for instance, Qiou and Ravishanker [ [28]. However, as elucidated by Samorodnitsky and Taqqu [ [32], sub-Gaussian models are ’not very far away from Gaussian models’ in the sense of Proposition 2.5.7 [32]. Moreover, not all symmetric stable laws are sub-Gaussian.

If individual return distribution of each item in a basket is likely to have thick tail, as most research with univariate laws suggests, it is more natural to use a multivariate law with this property. Due to generalised central limit theorem, multivariate stable laws are ideal candidates, but are also notorious for their mathematical intractability. Multivariate stable laws are implicitly described through its characteristic function which involves a measure on the unit sphere in Rd\mathbb{R}^{d}, known as spectral measure. Basic properties and tail dependence structures of multivariate stable laws are expressed in terms of the spectral measure.

Despite its importance in understanding multivariate stable laws, statistical estimation of spectral measure continues to be a major challenge. Attempts in estimating spectral measure so far have been based on a discrete approximation result by Byczkowski et al. [ [7] and results on univariate projections by Abdul-Hamid and Nolan [ [1]. Research by Nolan and his collaborators has given us reasonable understanding of bivariate stable laws; see, for instance, Nolan and Panorska [ [26] for bivariate stable densities in foreign exchange rates. Pivato and Seco [ [27] approached the problem of estimation of spectral measure through spherical harmonic analysis. More recently, Tsionas [ [35] developed Bayesian inference using one-dimensional projections of Matsui and Takemura [ [23].

In the present paper, we approach the problem using a LePage-type representation [20, 19] with the von Mises – Fisher distribution as the spectral measure. This approach is motivated by five factors.

von Mises through Brownian motion: Gordon and Hudson [ [12] showed that the von Mises distribution arises as the hitting distribution of a two-dimensional Brownian motion with drift; see also Gordon and Hudson [ [13]. The pole-seeking Brownian motion of Kendall [ [15] gives a vivid physical example through bird navigation. To our knowledge, the present work is the first to apply the Gordon–Hudson characterisation to motivate the spectral measure of a multivariate stable distribution. Gatto [ [11] extended the Gordon–Hudson result to higher dimensions.

Regular variation: For dd-dimensional stable laws, the tail probability and directional behaviour in any cone are jointly determined by the normalised spectral measure [Resnick, 1987, [3]. The von Mises – Fisher distribution provides an explicit parametric form for this measure, with the mean direction η\eta identifying the dominant tail direction and the concentration κ\kappa measuring tail anisotropy.

LePage-type representation: Damien et al. [ [10] developed an approximation theorem for simulating one-dimensional infinitely divisible distributions. We extend their approach to the multivariate setting, identifying a latent structure within stable laws — radial, spectral, and Poisson components — that yields a representation theorem in the spirit of Le Page et al. [ [19] and forms the basis of the Gibbs sampler (§ 3.2).

Truncation error through Gaussian: The LePage series is truncated at mm terms for numerical implementation, introducing an approximation error. We model this error through a Gaussian component, which yields an explicit likelihood function for all parameters jointly — a step that makes full Bayesian inference tractable for dd-dimensional stable distributions.

Radial component as latent variable: The classical approach integrates out the radial component analytically, yielding a characteristic function that requires special treatment at the Cauchy boundary α=1\alpha= 1. In the present paper, the radial component is retained as an explicit latent variable in the Gibbs sampler, following Kuelbs [ [18]. This has two consequences; the likelihood is smooth in α∈(0,2)\alpha\in(0, 2) with no special case at α=1\alpha= 1; and the radial draws carry direct information about α\alpha through a sufficient statistic, independently of the spectral parameters. This is a fundamental departure from existing approaches to multivariate stable inference.

Use of von Mises – Fisher distribution as a model for normalized spectral measure allows us to explore tools of directional statistics in an altogether different area of dd-dimensional stable laws. In particular, terminologies of directional statistics elicit natural and lucid interpretation of the mean direction of the spectral mass. Through von Mises – Fisher distribution, asymmetry of the spectral measure is controlled by its concentration parameter. We believe that this approach eliminates some of the limitations of discrete approximations and also seems to enhance the horizon of understanding of multivariate stable laws.

A perhaps surprising finding is that the Gaussian limit of the bivariate stable distribution with von Mises spectral measure has an explicit covariance matrix ΣΓ=Γ∞(I2+R2(κ)A(η))\Sigma_\Gamma= \Gamma_\infty(I_2 + R_2(\kappa)A(\eta)), where R2(κ)=I2(κ)/I0(κ)R_2(\kappa) = I_2(\kappa)/I_0(\kappa) is the second-order mean resultant length and Γ∞\Gamma_\infty is the total spectral mass. This connects the tail anisotropy of the stable distribution directly to the Gaussian covariance structure at the boundary. For general result in dd-dimension, see § 2.3.

The paper is organised as follows. Section 2 motivates the von Mises spectral measure through its connection to the hitting distribution of Brownian motion on the unit circle. Section 3 develops the LePage-type representation and complete-data likelihood. Identifiability of the stability index α\alpha is established in § 4. § 5 describes the Markov chain initialisation, with full details in the online Appendix. Section 6 presents the simulation design and results. Section 7 concludes.

Directional structure of multivariate stable distributions

Gordon–Hudson–Gatto Theorem

Gordon and Hudson [ [12] showed that the von Mises distribution arises as the hitting distribution of a drifted Brownian motion on S1S^1. Gatto [ [11] extended this result to general dimension d≥2d \ge2; the von Mises–Fisher distribution vMF⁡(μ,κ)\operatorname{vMF}(\mu,\kappa) on Sd−1S^{d-1} is the hitting distribution of a dd-dimensional Brownian motion with constant drift ν=κμ\nu= \kappa\mu on the unit sphere Sd−1S^{d-1}.

Theorem 1 (Gordon–Hudson–Gatto). Let {Bt}t≥0\{B_t\}_{t\geq0} be a dd-dimensional standard Brownian motion starting at the origin, with constant drift vector ν=κμ∈Rd\nu= \kappa\mu\in\mathbb{R}^d, where μ∈Sd−1\mu\in S^{d-1} and κ≥0\kappa\geq0. That is, Bt=Wt+κμtB_t = W_t + \kappa\mu t, with WtW_t a standard Brownian motion in Rd\mathbb{R}^d. Let τ=inf⁡{t>0:∥Bt∥=1}\tau= \inf\{t > 0 : \lVert B_t\rVert= 1\} be the first hitting time of Sd−1S^{d-1}, and let S=Bτ/∥Bτ∥∈Sd−1S = B_\tau/\lVert B_\tau\rVert\in S^{d-1} be the hitting direction. Then SS has the von Mises–Fisher distribution:

S∼vMF⁡(μ,κ),S \sim\operatorname{vMF}(\mu,\kappa),

with density

fvMF⁡(s∣μ,κ)=Cd(κ)exp⁡(κμTs),s∈Sd−1,(1)f_{\operatorname{vMF}}(s \mid\mu,\kappa) = C_d(\kappa)\exp(\kappa\mu^\mathsf{T}s), \qquad s \in S^{d-1}, \tag*{(1)}

where Cd(κ)=κd/2−1(2π)d/2Id/2−1(κ)C_d(\kappa) = \frac{\kappa^{d/2-1}}{(2\pi)^{d/2}I_{d/2-1}(\kappa)} is the normalising constant and Ip(⋅)I_p(\cdot) is the modified Bessel function of the first kind of order pp.

At d=2d = 2, (1) reduces to the von Mises density (2):

fvM⁡(θ∣η,κ)=eκcos⁡(θ−η)2πI0(κ),θ∈[0,2π),(2)f_{\operatorname{vM}}(\theta\mid\eta,\kappa) = \frac{e^{\kappa\cos(\theta-\eta)}}{2\pi I_0(\kappa)}, \qquad\theta\in[0,2\pi), \tag*{(2)}

where I0(κ)I_0(\kappa) is the modified Bessel function of the first kind of order zero.

This result motivates our choice of spectral measure for X∼Stable⁡(α,Γ)X \sim\operatorname{Stable}(\alpha,\Gamma) in Rd\mathbb{R}^d. Let Γ~=Γ/Γ∞\widetilde{\Gamma} = \Gamma/\Gamma_\infty be the normalised version of the spectral measure Γ\Gamma, where Γ∞=Γ(Sd−1)\Gamma_\infty= \Gamma(S^{d-1}). By setting Γ~∼vMF⁡(μ,κ)\widetilde{\Gamma} \sim\operatorname{vMF}(\mu,\kappa), the spectral measure acquires a natural Brownian motion interpretation—it is the distribution of exit directions of an independent, drifted Brownian motion from the unit sphere. The mean direction μ\mu identifies the dominant tail direction of the stable distribution and the concentration κ\kappa measures the degree of tail anisotropy. To the best of our knowledge, this connection between the vMF spectral measure and the drifted Brownian motion hitting distribution has not previously appeared in the stable distribution literature. Table 1 provides the parameter-level linkage for general dd.

ParameterInterpretation in stable modelInterpretation via BM hitting direction
μ∈Sd−1\mu\in S^{d-1}Mean direction of spectral mass concentrationDirection of BM drift vector κμ∈Rd\kappa\mu\in\mathbb{R}^{d}
κ≥0\kappa\ge0Concentration of spectral mass around μ\muMagnitude of drift ∥κμ∥\lVert\kappa\mu\rVert; larger κ\kappa means stronger pull toward μ\mu
κ=0\kappa= 0Uniform spectral measure on Sd−1S^{d-1}Zero drift: symmetric BM with uniform hitting direction
κ→∞\kappa\to\inftyDegenerate point mass at μ\muInfinite drift: BM hits the sphere deterministically at μ\mu
Rˉ=Id/2(κ)/Id/2−1(κ)\bar{R} = I_{d/2}(\kappa)/I_{d/2-1}(\kappa)Mean resultant length; measures degree of anisotropy (Rˉ=0\bar{R} = 0: isotropic; Rˉ→1\bar{R} \to1: degenerate)Expected projection of hitting direction onto μ\mu; larger Rˉ\bar{R} means more concentrated near μ\mu

Table 1. Linkage between the vMF spectral measure parameters and the first hitting direction of a dd-dimensional drifted Brownian motion on Sd−1S^{d-1} [12, 11].

Regular Variation Characterisation

The connection between the spectral measure Γ\Gamma and the tail behaviour of X∼Stable⁡(α,Γ)X \sim\operatorname{Stable}(\alpha,\Gamma) in Rd\mathbb{R}^{d} is made precise by the regular variation characterisation [30, 3]. With Γ~∼vMF⁡(μ,κ)\widetilde{\Gamma} \sim\operatorname{vMF}(\mu,\kappa), for any Borel set A⊂Sd−1A \subset S^{d-1}, Γ(∂A)=0\Gamma(\partial A) = 0, we have

lim⁡x→∞P(X/∥X∥∈A, ∥X∥>x)P(∥X∥>x)=Γ~(A)=∫AfvMF(s∣μ,κ) ds.(3)\lim_{x \to\infty} \frac{P(X/\lVert X\rVert\in A,\ \lVert X\rVert> x)}{P(\lVert X\rVert> x)} = \widetilde{\Gamma}(A) = \int_A f_{\mathrm{vMF}}(s \mid\mu,\kappa)\,\mathrm{d}s. \tag*{(3)}

Thus, conditional on lying in the far tail, the probability that XX points into direction set AA converges to the normalised spectral measure evaluated at AA. The spectral measure is therefore the limiting angular distribution of extreme observations. For the vMF spectral measure, (3) provides an explicit parametric form for this limiting distribution, valid in any dimension d≥2d \ge2, controlled by the mean direction μ\mu and concentration κ\kappa.

It is also worth mentioning that the von Mises–Fisher distribution vMF⁡(μ,κ)\operatorname{vMF}(\mu,\kappa) on Sd−1S^{d-1} is the unique maximum entropy distribution subject to a fixed mean resultant vector E[s]=sˉ\mathbb{E}[s] = \bar{s} (fixed mean direction μ∈Sd−1\mu\in S^{d-1} and fixed mean resultant length Rˉ=∥sˉ∥∈[0,1]\bar{R} = \lVert\bar{s}\rVert\in[0,1]) [22].

The Gaussian Boundary

At the Gaussian boundary, we show that all spectral measures induces a specific covariance structure. In the case of vMF, the covariance gets a closed form through its second moment matrix.

References

  1. [1]H. Abdul-Hamid and J.P. Nolan. Multivariate stable densities as functions of one dimensional projections. Journal of Multivariate Analysis, 67:80–89, 1998.
  2. [2]D. E. Amos. Computation of modified bessel functions and their ratios. Mathematics of Computation, 125:239–251, 1974.DOI
  3. [3]R. A. Basrak, B. Davis and T. Mikosch. A characterization of multivariate regular variation. The Annals of Applied Probability, 12:908–920, 2002.
  4. [4]D.J. Best and N.I. Fisher. The bias of the maximum likelihood estimators of the von Mises–Fisher concentration parameters. Communications in Statistics — Simulation and Computation, 10:493–502, 1981.DOI
  5. [5]O. Bradely and M.S. Taqqu. Financial risk and heavy tails. In S.T. Rachev, editor, Handbook of Heavy-Tailed Distributions in Finance. Elsevier, Amsterdam, 2003.
  6. [6]D.J. Buckle. Bayesian inference for stable distributions. Journal of the American Statistical Association, 90(430):605–613, 1995.DOI
  7. [7]T. Byczkowski, J.P. Nolan, and B. Rajput. Approximation of multidimensional stable densities. Journal of Multivariate Analysis, 46(1):13–31, 1993.DOI
  8. [8]P. Carr and L. Wu. The finite moment log stable process and option pricing. Journal of Finance, 58:753–777, 2003.
  9. [9]U. Cherubini, E. Luciano, and W. Vecchiato. Copula Methods in Finance. John Wiley, New York, 2004.DOI
  10. [10]P. Damien, P.W. Laud, and A.F.M. Smith. Approximate random variate generation from infinitely divisible distributions with applications to Bayesian inference. Journal of the Royal Statistical Society, Series B, 57:547–563, 1995.
  11. [11]R. Gatto. Information theoretic results for circular distributions. Statistics, 47(5):1008–1029, 2013.
  12. [12]L. Gordon and M. Hudson. A characterisation of the von Mises distribution. Annals of Statistics, 5:813–814, 1977.
  13. [13]L. Gordon and M. Hudson. Acknowledgment of priority. The Annals of Statistics, 7:927, 1979.
  14. [14]P. Jorion. Value at Risk: The New Benchmark for Managing Financial Risk. McGraw-Hill, New York, 3rd edition, 2007.
  15. [15]D. G. Kendall. Pole-seeking brownian motion and bird navigation. Journal of the Royal Statistical Society, Series B, 36:365–417, 1974.
  16. [16]Tarak Kharrat and Georgi N. Boshnakov. StableEstim: Estimate the Four Parameters of Stable Laws using Different Methods, 2024. URL https://CRAN.R-project.org/package=StableEstim. R package version 2.3.
  17. [17]I.A. Koutrouvelis. Regression-type estimation of the parameters of stable laws. Journal of the American Statistical Association, 75(372):918–928, 1980.DOI
  18. [18]J. Kuelbs. A representation theorem for symmetric stable processes and stable measures on H. Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete, 26:259–271, 1973.DOI
  19. [19]R. Le Page, M. Woodroofe, and J. Zinn. Convergence to a stable distribution via order statistics. Annals of Probability, 9(4):624–632, 1981.
  20. [20]Raoul LePage. Multidimensional infinitely divisible variables and processes. part i: Stable case. Technical Report 292, Department of Statistics, Stanford University, October 1980.
  21. [21]B. Mandelbrot. The variation of certain speculative prices. Journal of Business, 36:394–419, 1963.
  22. [22]K.V. Mardia and P.E. Jupp. Directional Statistics. Wiley, New York, 2000.DOI
  23. [23]M. Matsui and A. Takemura. Integral representations of one-dimensional projections for multivariate stable densities. Journal of Multivariate Analysis, 100:334–344, 2009.DOI
  24. [25]NIST. Digital Library of Mathematical Functions. Release 1.2.1 of 2024-06-15, 2024. URL https://dlmf.nist.gov/. F. W. J. Olver, A. B. Olde Daalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller, B. V. Saunders, H. S. Cohl, and M. A. McClain, eds.
  25. [26]J.P. Nolan and A.K. Panorska. Data analysis for heavy tailed multivariate samples. Communications in Statistics – Stochastic Models, 13:687–702, 1997.DOI
  26. [27]M. Pivato and L. Seco. Estimating the spectral measure of a multivariate stable distribution via spherical harmonic analysis. Journal of Multivariate Analysis, 87:219–240, 2003.DOI
  27. [28]Z. Qiou and N. Ravishanker. Bayesian inference for vector ARMA models with stable innovations. Sankhyā, Series A, 60:459–475, 1998.
  28. [29]S. T. Rachev and S. Mittnik. Stable Paretian Models in Finance. John Wiley, 2000.
  29. [30]S.I. Resnick. Extreme Values, Regular Variation, and Point Processes. Springer, 1987.
  30. [31]J. Rosiński. On series representations of infinitely divisible random vectors. Annals of Probability, 18(1):405–430, 1990.DOI
  31. [32]G. Samorodnitsky and M.S. Taqqu. Stable Non-Gaussian Random Processes. Chapman and Hall, New York, 1994.
  32. [33]J.Y. So. The distribution of foreign exchange price changes: trading day effects and risk measurement — a comment. Journal of Finance, 42:181–188, 1987.DOI
  33. [34]E.G. Tsionas. Monte Carlo inference in econometric models with symmetric stable disturbances. Journal of Econometrics, 88:365–401, 1999.DOI
  34. [35]M. G. Tsionas. Bayesian analysis of multivariate stable distributions using one-dimensional projections. Journal of Multivariate Analysis, 143:185–193, 2016.
  35. [36]V.M. Zolotarev. One-Dimensional Stable Distributions, volume 65 of Translations of Mathematical Monographs. American Mathematical Society, Providence, Rhode Island, 1986.

Paper details

vlnc:W_we86zgju
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

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

  • Mathematics

Claude Sonnet 4.6 (bare)

View statement
Looking for related papers…
Contents