Julien Masselot, Principal and Head of Capital and Risk at Barnett Waddingham (BW), explains how the consultancy's quantitative analysis tool, validateR, speeds up internal model investigations, while making the validation process more accessible to non-actuaries.

What are the latest trends in internal model validation and how is it evolving? Is it evolving enough?

The last decade has seen insurance capital modelling and validation teams making huge progress in model development, analysis of change, model use and model validation. However, the size of such teams has not grown with the demands imposed upon them. As such, they find themselves constantly chasing the next delivery, often without time to stop and think about what the next evolutionary step in their work should be.

We generally agree the validation process in itself, whether performed by first line or second line teams, has not changed significantly in the years of Solvency II. These validation processes may have been updated along the way to keep up with market practice, but are still largely in the same form as the early days of Solvency II. Processes tend to be manual, repetitive and relatively slow compared to the pace at which capital modelling software has improved over time.

While validation is an integral part of the risk and regulatory framework around internal models, the approach needs to evolve to cope with the fast-changing (internal and external) environment. This will then mean valuable feedback can be provided to both the first line and the board.

Certainly, from our experience, firms are beginning to ask whether the validation process is still fit for purpose.

How will validateR speed up the interpretation and analysis of internal models?

validateR has been designed to speed up the interpretation and analysis of internal model results through automation, while making the validation process more accessible to non-actuaries. This is how it works:

• The user obtains the simulated output from a stochastic model in a raw templated form. Once the initial set up has been agreed with the capital team, subsequent data extraction can be automated as part of standard model change processes.
• This template is then loaded into the validateR app. validateR is cloud based and so can be used on capital models of all sizes.
• Users can then use a wide array of tests and tools to analyse the model. It will also generate a comprehensive analysis of change as well as industry-standard metrics for benchmarking.
• At the click of a button, the user can execute the chosen tests and download a suite of analysis spreadsheets which include visualisations in Microsoft Excel in a repeatable and structured process which is accessible to all members of the team, even those who are new to stochastic modelling or who are not technically confident.

Does BW plan to further develop validateR this year?

Our goal is to add further capability to insurers' validation and capital teams throughout the course of the year, allowing them to move beyond performing validation only as part of their annual capital assessment.

Our development plans are influenced by users' feedback to prioritise those tests they will find most impactful first. In the longer term, we are exploring extending the tool to validate other stochastic-based models used by insurers.

How do you expect internal model validation to change over the next decade? Will technology have a role?

Technology has had, and will continue to have, a crucial role to play in automating many aspects of model validation. Increasing efficiency and the value-add aspect of the validation process has been a key driver in the deployment of tools to analyse the large amount of data and parameters feeding into capital models. These provide live validation feedback to first-line capital teams and allow for faster adjustments and improvements.

However, we need to be mindful that some of the most insightful challenges provided by validation teams usually does not relate to the calibration process, but rather the methodology used and relevance of the company's risk profile.

There is also an opportunity to leverage artificial intelligence to provide and update documentation, policies and internal model reporting within an insurance company, going beyond the initial remit of the capital model, but then supporting the wider business and ensuring consistency in decision-making.

We are immensely proud and honoured to receive the prestigious InsuranceERM Risk Innovation of the Year award. We look forward to empowering more insurers by transforming validation into a powerful, value-adding tool within the capital modelling process.

For more information or to book a demo, please see our website: https://www.barnett-waddingham.co.uk/services/validater/