[Paper Review] A General Asymptotic Implied Volatility for Stochastic Volatility Models

Pierre Henry‐Labordère|arXiv (Cornell University)|Apr 13, 2005

Stochastic processes and financial applications10 references4 citations

TL;DR

This paper develops a general asymptotic formula for implied volatility in stochastic volatility models using heat kernel expansion on a Riemannian manifold with an Abelian connection. It derives a first-order asymptotic implied volatility that improves upon the Hagan et al. formula for the SABR model and provides exact solutions for β=0 and β=1 cases, establishing a geometric link between the Poincaré hyperbolic plane and the λ-SABR model.

ABSTRACT

In this paper, we derive a general asymptotic implied volatility at the first-order for any stochastic volatility model using the heat kernel expansion on a Riemann manifold endowed with an Abelian connection. This formula is particularly useful for the calibration procedure. As an application, we obtain an asymptotic smile for a SABR model with a mean-reversion term, called lambda-SABR, corresponding in our geometric framework to the Poincaré hyperbolic plane. When the lambda-SABR model degenerates into the SABR-model, we show that our asymptotic implied volatility is a better approximation than the classical Hagan-al expression . Furthermore, in order to show the strength of this geometric framework, we give an exact solution of the SABR model with beta=0 or 1. In a next paper, we will show how our method can be applied in other contexts such as the derivation of an asymptotic implied volatility for a Libor market model with a stochastic volatility.

Motivation & Objective

• To derive a general first-order asymptotic implied volatility formula applicable to any stochastic volatility model.
• To unify the calibration of stochastic volatility models through geometric methods based on Riemannian manifolds and Abelian connections.
• To improve upon the classical Hagan-al asymptotic formula for the SABR model by incorporating a mean-reversion term (λ-SABR).
• To provide exact solutions for the conditional probability in the SABR model when β=0 or β=1.
• To establish a geometric correspondence between the Poincaré hyperbolic plane and the λ-SABR model.

Proposed method

• Utilizes heat kernel expansion on a Riemannian manifold to derive the short-time asymptotic solution of the backward Kolmogorov equation for the conditional probability density.
• Expresses the local volatility function as the mean of the stochastic volatility, derived via a covariant formulation of the Fokker-Planck equation.
• Applies the saddle-point method to obtain a first-order asymptotic expansion of the local volatility function in terms of the manifold's geometric structure.
• Introduces an Abelian connection on a line bundle to model the drift processes, which contributes to the first-order correction in the implied volatility expansion.
• Establishes a one-to-one asymptotic correspondence between local volatility and implied volatility using heat kernel methods on a time-dependent real line.
• Applies the Laplace method for multi-dimensional integrals to compute the leading-order and first-order asymptotics of the option pricing integral.

Experimental results

Research questions

• RQ1 How can a general first-order asymptotic implied volatility formula be derived for arbitrary stochastic volatility models?
• RQ2 What is the geometric interpretation of the zero-order and first-order terms in the implied volatility expansion?
• RQ3 How does the inclusion of a mean-reversion parameter (λ) in the SABR model affect the asymptotic smile compared to the classical SABR model?
• RQ4 Can exact solutions be derived for the conditional probability in the SABR model when β=0 or β=1?
• RQ5 What is the role of the Abelian connection in capturing the drift-dependent first-order correction in the implied volatility?

Key findings

• The zero-order asymptotic implied volatility is determined by the geodesic distance on the Riemannian manifold, linking the geometric structure to the implied volatility smile.
• The first-order correction in the implied volatility expansion depends on an Abelian connection that encodes the drift processes of the underlying stochastic volatility model.
• For the λ-SABR model, the derived asymptotic implied volatility provides a better approximation than the classical Hagan-al formula, especially in capturing the correct skew behavior.
• Exact solutions for the conditional probability are obtained in the SABR model when β=0 or β=1, confirming the consistency of the geometric framework.
• The geometric framework establishes that the λ-SABR model corresponds to the Poincaré hyperbolic plane, providing a rigorous differential-geometric interpretation of the model.
• The method enables a systematic derivation of asymptotic implied volatilities for other models, such as the Libor market model with stochastic volatility, as outlined in future work.