Asymptotic chaos expansions in finance. Theory and practice (Q2250287)

This book presents the recently developed Asymptotic Chaos Expansions (ACE) approach, which addresses to closed form valuation for European options in Stochastic Volatility model classes. The ACE algorithm links the Stochastic Instantaneous Volatility (SInsV) model and Stochastic Implied Volatility (SImpV) model classes by combining standard partial differential equations and stochastic differential equations approaches, to provide pure asymptotic of the smile's shape and dynamics. This book consists of 8 chapters and appendices. Chapter 1 is an Introduction and it contains review of stochastic volatility market models, asymptotic methods and relevant literature. Part 1 (``Single underlying'') consists of Chapters 2, 3 and 4. In Chapter 2 the author presents the core principles of the ACE methodology. It is investigated the relationship between stochastic instantaneous volatility and stochastic implied volatility models, in the simple case of a single underlying, and when the endogenous driver is scalar. In Section 2.1 it is defined the market environment: the underlying, the numeraire and the liquid European options. The author defines the re-parametrisation of the option price surface via a sliding implied volatility map. It is stated the author's objectives, which it is split into a direct and an inverse problem. In Section 2.2 the author establishes the Zero Drift Condition, a partial differential equations constraining the shape and dynamics of the stochastic implied volatility model in the whole strike/expiry domain, in order to respect the no-arbitrage assumption. In Section 2.3 it is solved part of the inverse problem, which is to recover the instantaneous model from the implied one. The author shows that at a given level of precision (the first layer, which involves a group of low-order differentials of the smile) the implied model injects itself into instantaneous class. Section 2.4 deals with direct problem, which consists in generating the smile, and more generally the implied model associated to a given instantaneous class. For the first layer, it is established the opposite connection from before. Section 2.5 is devoted to some applications of obtained results. In term of pure asymptotic, the author defines a stochastic instantaneous volatility model class, covering most popular stochastic volatility models, for which it is provided the first layer differentials. Section 2.6 contains the conclusions of Chapter 2. Chapter 3 is dedicated to generalisation of the ACE approach and its results. In Section 3.1 it is extended the results of Chapter 2 concerning the direct problem to higher differentiation orders. The author proves that any pure or cross-differential of the smile shape and dynamic maps, taken at the Immediate At The Money (IATM) point, can be computed from the dynamics of a generic stochastic instantaneous volatility model. The ladder effect which specifies structurally the order of the computations and organise the differentials into well-defined groups called layers is considered. In Section 3.2 the author illustrates the presented generic methodology by computing the second and third layers. In Section 3.3 it is decomposed the overall methodology into independent building blocks, and it is reviewed how each of these can be generalised. The author justifies and examines the introduction of alternative baselines to replace the lognormal model. A generic but numerical method that allows us to transfer the results from one existing baseline to the next is presented. Section 3.4 is devoted to extension of Chapter 2 results on the first layer to multi-dimensional underlyings, drivers and dynamic coefficients. Section 3.5 deals with illustrations of the previous multi-dimensional results in the context of baskets. The author considers the case of general stochastic weights as well as of fixed weights, with scalar and with tensorial individuals. Chapter 4 deals with practical considerations, by applying ACE results to some popular stochastic instantaneous volatility models, namely the Stochastic Alpha Beta Rho (SABR) and Forward Libor -- Stochastic Volatility (FL-SV) classes. In Section 4.1 it is discussed the practicality and expected performance of a Taylor expansion as a smile proxy, destined to be used for calibration and/or within a numerical scheme, or for hedging. Then in Section 4.2 the author applies these principles to the generic SABR model, and derives the coefficients of the chaos diffusion which are required by all static IATM differentials. Then it is used some of these coefficients to establish bilateral relationships between the model parameter and the IATM level, skew and curvature. In Section 4.3 the author obtains the relevant coefficients of the chaos diffusion, links explicitly the model parameters to the IATM smile, and shows that Hagan's formula does verify these IATM differentials. Section 4.4 deals with stochastic instantaneous volatility model, the FL-SV class. It is introduced a more generic version of the class, called Extended FL-SV. The the author presents as an exercises the computation of all static IATM differentials within the first layer. Solutions of these exercises are presented at the end of the book. In Section 4.5 the author illustrates obtained results with a few simulations. Part 2 (``Term structures'') of the book consists of Chapters 5--8, and deals with a term structure framework. Chapter 5 is devoted to a generalised framework dealing with a collection of underlyings. Section 5.1 presents the framework by considering the three families of processes defining term structure framework: underlyings, numeraires and European call options. Then it is considered the implied and instantaneous stochastic volatility models. In Section 5.2 it is expressed the Main Zero-Drift Condition (ZDC), defined regularity assumptions and developed the Immediate ZDC. In Section 5.3 the author establishes the recovery theorem, which deals with solving the inverse problem. In Section 5.4 it is solved the direct problem, which is intensively invoked for calibration purposes. The first layer of IATM differentials for the smile is derived. In the Section 5.5 further research and applicability of the method are discussed. Chapter 6 deals with very liquid interest rates derivatives products, valued within a universal Stochastic Volatility Heath-Jarrow-Morton (SV-HJM) modelling setup. In Section 6.1 the author sets the input SV-HJM modelling framework in a generic, chaos diffusion style. He defines and justifies his objectives in terms of the specific derivative products targeted, as well as the required precision of the output smile shapes and dynamics. In Section 6.2 some useful intermediate results, pertaining to the dynamics of rebased bonds, first single zero-coupons and then full-blown fixed structures are presented. These results allows author to describe the chaos diffusion of a driftless process. Section 6.3 deals with the first type of derivative product, the European call options on fixed-coupon bonds. The author immerses the problem into the generic setup, and then computes accordingly the chaos dynamics for the chosen underlying. In Section 6.4 the author investigates the caplets, expressing all the \(\sigma\)-(2,0) coefficients necessary to formulate the \(\tilde\Sigma\)-(2,0) group of differentials. Section 6.5 similarly covers the European payer swaptions. In Section 6.6 it is discussed the possibility of obtaining the same price surface information through an indirect approach, by considering options on assets rather than on the corresponding rates. This approach concerns caplets and swaptions, which it is treated separately. In Chapter 7 the author applies the generic term structure framework defined in Chapter 5 to the particular case of interest rates options, within a universal Stochastic Volatility Libor Market Model (SV-LMM). Section 7.1 presents the generic LMM modelling framework, including its specific chaos dynamics. In Section 7.2 it is expressed the dynamics of the rebased zero-coupons with the LMM parametrisation. Section 7.3 deals with the first liquid product defining an implied volatility surface, the bond option. The author immerses the problem into the generic term structure framework, and then derives the chaos dynamics. Section 7.4 deals with the simplest case, which is the caplet smile. Since the chaos dynamics are both given and simple, the author provides some interpretation w.r.t. the IATM level, skew and maturity slope. In Section 7.5 it is studied the physical swaptions. The goal of this section is to link the SInsV specification of the generic SV-LMM class to the statics and dynamics of the swaption implied volatility surface. In particular, the author is interested in the IATM differentials of that smile. which corresponds to the most liquid options in the interest rates market. Section 7.6 deals with the usual approximation for swaptions' volatility, which employs a fixed-weight basket of Libor rates as a proxy underlying. It is expressed the exact and multi-dimensional error generated on the weighting scheme, then it is specialised this result to an idealised affine Libor curve setting. A numerical example is presented. Chapter 8 consists of conclusions. The five appendices cover the Itô-Kunita formula, transition formulae, Black and Bachelier differentials, the linear algebra toolbox, and the computation of the 2nd and 3rd layers.