Paper: arXiv 2307.13870
Authors: Andrey Itkin, Dmitry Muravey
Abstract
Semi-analytical pricing of American options in a time-dependent Ornstein-Uhlenbeck model was presented in [“Carr, Itkin, 2020”]. It was shown that to obtain these prices one needs to solve (numerically) a nonlinear Volterra integral equation of the second kind to find the exercise boundary (which is a function of the time only). Once this is done, the option prices follow. It was also shown that computationally this method is as efficient as the forward finite difference solver while providing better accuracy and stability. Later this approach called “the Generalized Integral transform” method has been significantly extended by the authors (also, in cooperation with Peter Carr and Alex Lipton) to various time-dependent one factor, and stochastic volatility models as applied to pricing barrier options. However, for American options, despite possible, this was not explicitly reported anywhere. In this paper our goal is to fill this gap and also discuss which numerical method (including those in machine learning) could be efficient to solve the corresponding Volterra integral equations.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 4.0/10
- Quadrant: Lab Rats — high math complexity, limited empirical testing
Why this score: The paper is heavily theoretical, deriving nonlinear Volterra integral equations for exercise boundaries across multiple models (OU, Hull-White, CEV, CIR) and discussing analytical transforms and numerical solvers, indicating high mathematical density. However, it contains no empirical backtests, code, or data; the focus is on algorithmic formulation and numerical efficiency, placing it in the ‘Lab Rats’ quadrant.
Research Flowchart
flowchart TD A["Research Goal<br>Extend semi-analytical pricing<br>for American options in time-dependent<br>Ornstein-Uhlenbeck models<br>and evaluate efficient numerical solvers"] --> B["Methodology<br>Generalized Integral Transform"] B --> C["Mathematical Formulation<br>Derive Volterra Integral Equation<br>for the exercise boundary S(t)"] C --> D["Data & Model Inputs<br>• Model parameters (OU process)<br>• Option specs (T, K, r)<br>• Initial asset price S0"] D --> E["Computational Process<br>Solve Volterra Equation for S(t)<br><br>Numerical Methods:<br>• ML-supported solvers<br>• Numerical quadrature<br>• Finite difference schemes"] E --> F["Outcomes<br>• American option price P(t,S)<br>• Exercise boundary S(t)<br>• Computational efficiency &<br>accuracy comparison vs. PDE solvers"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper extends the Generalized Integral Transform (GIT) technique to semi‑analytic pricing of American options and discusses numerical (including ML) methods to solve the resulting nonlinear Volterra integral equations for the exercise boundary.
Source sentence
In this paper we demonstrate how the GIT technique could be extended from semi-analytic pricing of barrier options, [Itkin et al., 2021] to that for American options.
Assumes 4
- The exercise boundary FB(t) (or SB(t)) is known for the derivations. model
Source sentence
Using the same argument as in Section 1.1, let’s assume that the exercise boundary FB(t) is known.
- All model parameters are known as continuous functions of time or discrete values. model
Source sentence
We assume that all parameters of the model are known either as a continuous functions of time t ∈[0, ∞), or as a discrete set of N values for some moments ti, i = 1, . . . , N.
- The pricing PDE can be reduced to the heat equation (or Bessel equation) with a source term and moving boundaries. model
Source sentence
In this section we consider two typical examples of one-factor time-dependent models such that the American option pricing can be reduced to solving the heat equation at the time-dependent spatial domain.
- The underlying stochastic processes follow the specified time‑dependent Ornstein‑Uhlenbeck, Hull‑White, Verhulst, CEV, or CIR models. model
Source sentence
... pricing of the American Put option written on a zero‑coupon bond where the underlying interest rate follows the time‑dependent Hull‑White model.
Data
Synthetic data only.
Costs
Not applicable
Finds 3
- The GIT method is computationally as efficient as the forward finite‑difference solver while providing better accuracy and stability for American option pricing. theoretical
Source sentence
It was also shown that computationally this method is as efficient as the forward finite difference solver while providing better accuracy and stability.
- Numerical experiments show that the elapsed time for the semi‑analytical GIT method is comparable to that of a trinomial tree (e.g., 0.05 sec per strike). performance
Source sentence
the elapsed time in both tests is very close, despite here we used our semi-analytical method, and for the test in Fig. 1 - the trinomial tree.
- All Greeks for American options can be obtained in one sweep with the GIT method, unlike finite‑difference methods that require separate runs. theoretical
Source sentence
For the FD method, the Greeks can be found only numerically. Moreover, to compute, e.g. the option Vega, a new run of the FD method is required, while for the GIT method all Greeks can be calculated in one sweep.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 21 related papers). Coverage is partial, so this is neither replication nor refutation.
6 papers in the archive
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.