When Multi-Factor Option Models Need Numerical Methods
Summary
The document asks when a European option pricing problem based on the Black–Scholes equation cannot be solved analytically. The response explains that the standard model has one source of uncertainty, while extensions can introduce additional risk factors, such as stochastic interest rates or stochastic volatility. These extra state variables can make the resulting pricing equation harder to reduce to a constant coefficient diffusion equation, leading practitioners to consider numerical methods such as finite differences.
The response uses the Heston stochastic volatility model as an example of a multi-factor extension, while noting that it can still be priced without finite differences. Thus, adding a factor does not automatically rule out analytical or other non-grid solutions; the tractability depends on the specific model. The exchange gives no worked derivation or concrete example of a model that necessarily requires numerical treatment, and points only generally toward further study. Its main lesson is that solvability is model dependent rather than determined simply by whether an option is European or American.
Key ideas
- The standard Black–Scholes model has one source of uncertainty, while extensions can add risk factors.
- Stochastic interest rates and stochastic volatility are examples of additional factors.
- The Heston model includes stochastic volatility but can be priced without finite differences.
- Whether a model has an analytical solution depends on its structure, not only on the option’s exercise style.
- The document offers no derivation of a model that requires numerical methods.
Tags
Full text
# When is a numerical solution the only way to obtain a solution to BS? # When is a numerical solution the only way to obtain a solution to BS? I am only now reading into Mathematical Finance, I understand the derivation of the BS equation with vanilla European options. On the next page of my book it starts to delve into obtaining exact solutions for the BS equation for Euro-options, and the introductory chapter on numerical methods has this to say: > [...] There are many examples (particularly of multi-factor models) where it is not feasible or even not possible to reduce the problem to a constant coefficient diffusion equation; in this case there is little choice but to use finite differences on the BS equation [...]" There are no links to these models or relevant chapters in the book. I have googled "multi-factor black scholes" and I am not getting anything digestible. Question: Could someone take me through an instance of the BS equation using European options that cannot be solved analytically? Possible some references to derivation? On first thought I thought it was something to do with the type of option (European over American) but it seems as though you can get solutions to American options too. ## Answer by Evgenii (score 2, accepted) https://quant.stackexchange.com/a/47540 This is my first answer here in the StackExchange. In the standard BS equatation the uncertainty is driven purely by the Brownian shocks and therefore it is a single risk factor model. However, it is possible to extend this model and add another risk factors and you will get a multi-factor model for pricing options. For example, you can do that by introducing uncertainty into the interest rates or volatility process. One of the widely used multi-factor models is the so-called Heston model, where volatility process is additionally modeled as a stochastic process. See https://en.wikipedia.org/wiki/Heston_model. In the Heston model you can also price option without a finite difference method, but there are more complicated multi-factor models that you would not be able to solve without numerical methods.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.