Risk Modelling and Validation

Models are central to risk and pricing environments. We have experience of building, selecting and validating models in a wide variety of areas.

Within these areas we provide a range of services, described below.

Model development

We can write customised pricing and risk models for general or for specific derivative instruments and structures. The models may produce theoretical values, first and second order greeks, VaR, potential future exposure, CVA, etc. Models can be written in C, C++, C# or VBA. Clients are provided not only with a working solution but also the code in the form of a buildable project, extensive documentation and a full test suite.

Model package selection

Many clients choose to buy an “off-the-shelf” package to solve their modelling requirements rather than building in-house. We have in-depth knowledge of model vendors. We do not believe there is a one-size fits all best package. The most suitable package for a client depends on the products they wish to cover, the potential extensibility of both the models and their business, their sophistication level and how they will use the results of the models.

See our System Selection and Solution Definition services for our wider capabilities and approach in the area of system selection.

Model implementation

As derivatives get more complicated and the research of financial mathematics gets more complex, solutions to modelling requirements can also be extremely complicated. In all but the most simple cases, the days of turn-key modelling solutions are gone. Most model packages require an investment in time and expertise to get the best out of them. 

Model choice best practice

As the boundaries of financial mathematics have been further expanded, different methods can now be used to model the same structure. Take, as an example, a simple barrier option. One could price this using (i) single-volatility Black-Scholes, (ii) two-volatility Black-Scholes, (iii) local volatility, (iv) Merton jump diffusion, (v) stochastic volatility etc. There is no such thing as the right model for a product. The most suitable model of a product depends on the clients’ business: why they are pricing the product, what risks are they are aiming to identify and capture, and how they might hedge the product. We can advise on best market practice in the context of their business – helping choose the models which will provide the correct blend of sophistication and efficiency.

Model validation

As regulatory regimes regarding derivatives become more stringent, the validation of models being used by firms has become a key area for scrutiny. We can provide independent, thorough and fully-documented validation of models and the implementation of such models.

We have a rigorous methodology for the validation of models. Alternatively, we can adhere to clients’ own model validation policies and standards. 

Clients may wish to engage us on an ongoing basis to act as their validation provider (as part of our Outsourced Risk Support Services offering), or on a one-off basis to validate specific models.