European Options in Market Models with Multiple Defaults: the BSDE approach

Miryana Grigorova, James Wheeldon · 2026

Read European Options in Market Models with Multiple Defaults: the BSDE approach by Miryana Grigorova, James Wheeldon (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-01250.txt.

Introduction

Miryana Grigorova and James Wheeldon price European options in markets where several assets can default. They use backward stochastic differential equations, a tool that works backward from an option's payoff at expiry to its price today, driven by both smooth market motion and sudden default jumps. Their setup allows a general driver term with an added finite-variation process.

First, they prove that solutions exist and are unique, then establish comparison results that order prices when inputs are ordered. For a linear driver, they derive an explicit formula for the price component. The formula takes two forms depending on how predictable the added process is. They then apply the theory to hedging in complete markets with defaultable assets, including cases where a large seller's own trades shift the odds of default. The work gives traders and theorists a firm base for pricing under default risk.

Catalog record

CREATORMiryana Grigorova, James Wheeldon
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE117 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-01250

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

European Options in Market Models with Multiple Defaults: the BSDE approach Miryana Grigorova, James Wheeldon Abstract We study non-linear Backward Stochastic Differential Equations (BSDEs) driven by a Brownian motion and p p default martingales. The driver of the BSDE with multiple default jumps can take a generalized form involving an optional finite variation process. We first show existence and uniqueness. We then establish comparison and strict comparison results for these BSDEs, under a suitable assumption on the driver. In the case of a linear driver, we derive an explicit formula for the first component of the BSDE using an adjoint exponential semimartingale. The representation depends on whether the finite variation process is predictable or only optional. We apply our results to the problem of pricing and hedging a European option in a linear complete market with two defaultable assets and in a non-linear complete market with p p defaultable assets. Two examples of the latter market model are provided: an example where the seller of the option is a large investor influencing the probability of default of a single asset and an example where the large seller’s strategy affects the default probabilities of all p p assets. Keywords: BSDEs with Multiple Default Jumps, Generalized Driver, Comparison Theorem, European Option, Non-linear Market, Optional Dividend Process. 1 Introduction In this paper, we consider BSDEs with multiple default jumps and explore some applications in a financial context. BSDEs were first introduced by Bismut [ 4 ] , who studied the linear case in relation to stochastic control. Pardoux and Peng [ 19 ] established the well-posedness of non-linear BSDEs with Lipschitz drivers in a Brownian filtration. BSDEs incorporating both continuous and jump components have also been studied in the literature (cf., e.g., [ 21 ] , [ 20 ] , [ 1 ] , [ 7 ] , [ 18 ] ). In [ 7 ] , BSDEs with a single default jump are studied and applied in a financial context (for developments in the incomplete single default framework, we refer to [ 14 , 15 ] ). In the present paper, we consider BSDEs driven by a Brownian motion and p compensated default martingales, each associated with a default time \tau_{i}>0 and an intensity process \lambda^{i}=(\lambda_{t}^{i})_{t\in[0,T]} for i\in\{1,\ldots,p\} . We develop new a priori estimates and establish the existence and uniqueness of the solution to the BSDE with multiple default jumps. We consider an optional (not necessarily predictable), right-continuous left-limited (rcll) process D of finite variation which enters into the driver of the BSDE, leading to a generalized form of the driver: g(t,y,z,k^{1},\ldots,k^{p})dt+dD_{t} . This modeling choice is motivated by the fact that, in markets with defaultable securities, contingent claims often give rise to intermediate cash flows, particularly at the time of default, as observed in [ 2 ] . We extend the definition of \lambda -linear drivers introduced in [ 7 ] to the general case of \lambda^{(p)} -linear drivers, where \lambda^{(p)} refers to the vector of p intensity processes \lambda^{1},\ldots,\lambda^{p} . When the driver g is \lambda^{(p)} -linear, we derive an explicit representation of the solution to the BSDE with multiple default jumps using an associated adjoint semimartingale. This representation depends on whether the process D is predictable or only optional. We also prove a comparison result and a strict comparison result under suitable assumptions on the driver, where we distinguish again the case where D is predictable and the case where D is only optional. We present two financial applications of this model: one in which the market is linear and complete, and another in which the market is complete but non-linear. In both cases, we assume the existence of a risk-free asset, a default-free (jump-free) risky asset, and two or more defaultable assets. We focus on pricing and hedging a European contingent claim with terminal payoff \eta at maturity T>0 , and with intermediate ‘dividends’ modeled by the process D , where D represents an exogenous cumulative process. The process D is not necessarily predictable. In the second example, the market is non-linear, with the non-linearity arising from imperfections caused by a large seller whose strategy influences default probabilities. In the non-linear market setting with p defaults, we show that the price of the contingent claim is X_{\cdot,T}^{g}(\eta,D) , where X_{\cdot,T}^{g}(\eta,D) denotes the solution to the non-linear BSDE with multiple default jumps, terminal time T , terminal condition \eta , and generalized driver of the form g(t,y,z,k^{1},\ldots,k^{p})dt+dD_{t} . This gives rise to a non-linear pricing system \mathbf{X}^{g}:(\eta,D)\mapsto X_{\cdot,T}^{g}(\eta,D) , whose properties we study. When D is fixed, we define the associated (g,D) -conditional evaluation by \mathscr{E}_{\cdot,T}^{g,D}(\eta)\coloneqq X_{\cdot,T}^{g}(\eta,D) for \eta\in L^{2}(\mathcal{G}_{T}), and provide its main properties. The remainder of the paper is organized as follows: Section 2 is dedicated to the study of the BSDE with p default jumps and generalized driver. In Subsection 2.1 , we define the BSDE with p default jumps and generalized \lambda^{(p)} -driver (allowing for an intermediate optional finite variational process D ); in Subsection 2.2 , we prove existence and uniqueness, and study the case where the generalized \lambda^{(p)} -driver is linear; in Subsection 2.3 , we establish comparison and strict comparison results under suitable assumptions on the driver. Section 3 is dedicated to the applications to pricing and hedging a European option in complete linear and non-linear markets. In Subsection 3.1 , we present an example of a linear and complete market; in Subsection 3.2 , we present the non-linear complete market with p defaults, and study properties of the associated non-linear pricing system for a European option; in Subsection 3.3 , we provide a particular case of a non-linear complete market where a large seller influences the probability of default of one (defaultable) risky asset and, in Subsection 3.4 , we consider the example where the large seller’s trading strategy influences the probabilities of default of all p (defaultable) assets. 2 The Underlying Probability Setup In the sequel, we fix T>0 to be the finite time horizon. Let (\Omega,\mathcal{F},P) be a complete probability space, and let (W_{t})_{t\in[0,T]} be a one-dimensional Brownian motion. Let \mathbb{F}\coloneqq(\mathcal{F}_{t})_{t\in[0,T]} denote the augmented filtration generated by W . Inequalities and equalities between random variables are to be understood in the almost sure (a.s.) sense with respect to P . Let p\in\mathbb{N}\setminus\{0\} , and let \tau_{1},\tau_{2},\ldots,\tau_{p} be positive random times. We assume that, for each i\in\{1,\ldots,p\} , the random time \tau_{i} has a continuous distribution, and that P(\tau_{i}\neq\tau_{j})=1 for all i,j\in\{1,\ldots,p\} with i\neq j . Moreover, we assume that the p default times are strictly ordered, i.e., \tau_{1}<\tau_{2}<\cdots<\tau_{p} . We will interpret \tau_{i} as the i -th default (or i -th credit event) time. For i\in\{1,\ldots,p\} and t\in[0,T] , we define N_{t}^{i}\coloneqq\mathbbm{1}_{\{\tau_{i}\leq t\}} . For each i\in\{1,\ldots,p\} , let \mathbb{F}^{i}\coloneqq(\mathcal{F}_{t}^{i})_{t\in[0,T]} denote the smallest right-continuous filtration making \tau_{i} an \mathbb{F}^{i} -stopping time. We define the enlarged filtration \mathbb{G}\coloneqq\mathbb{F}\vee\mathbb{F}^{1}\vee\cdots\vee\mathbb{F}^{p} . The filtration \mathbb{G} is, in fact, the augmented filtration generated by W and the default indicator processes N^{1},N^{2},\ldots,N^{p} . We assume that the \mathbb{F} -Brownian motion W remains a \mathbb{G} -Brownian motion. By definition, for each i , the process N^{i}=(N_{t}^{i})_{t\in[0,T]} is non-decreasing, \mathbb{G} -adapted, and thus a \mathbb{G} -submartingale. Let \Lambda^{i}=(\Lambda_{t}^{i}) denote the \mathbb{G} -predictable compensator of N^{i} .

Read the complete text