Cash flow model migration

How we took model runs from eight hours to five minutes

Christelle Oosthuizen

Image by AI on Freepik

In a nutshell

Our Life team (Insight Life Solutions) worked with a growing life insurer in South Africa that offers a standard suite of long-term insurance products. We successfully migrated the insurer’s existing cash flow models from Excel to a single Mo.net model, which is now used for financial and regulatory reporting.

We began working with the client in June 2023 and used Excel for the June valuation. One of our analysts developed a single Mo.net model for all seven products that the life insurer has to offer, and it was used for the September valuation.

Due to Insight’s proprietary modelling paradigm and assumption system, the new model produces results more quickly and reliably.

 

How we did it?

  • One analyst built and fully tested all models in three hundred hours, allowing for its use in the September valuation.
  • Results from the Excel models were perfectly replicated in Mo.net.
  • We easily demonstrated the new model to the client, completing a test run almost instantaneously.
  • We developed a version control system that can be used to show the impacts of basis changes, refinements to the model, etc.

 

Benefits of the solution

  • A run that took nearly eight hours in Excel now takes just over five minutes in Mo.net using the same hardware.
  •  The new model directly integrates data from SQL, automating the data process and further streamlining the client’s valuation process.
  • Unlike Excel, Mo.net offers versatile ways to group outputs ready for use.
  • Users can initiate multiple tasks with a single click using task code.
  • Assumptions can be stored in an Excel file and linked directly to Mo.net, which enables easy assumption changes.
  • Models are created with VB.NET, a widely used and easily understood programming language with many online coding resources.

 

 

Insight’s proprietary modelling paradigm

We built a modular environment in Mo.net with a building block approach. This allows for easy code reuse when implementing new benefits and features.

 

  • Our models are built with a focus on parameter and data-driven modelling. This means that a wide range of product features can be modelled by just changing the inputs without having to change any code.
  • Lastly, we have invented the idea of a marketing and an actuarial layer in the code. The marketing layer is a series of sub-models that map straight back onto a product spec (in terms of naming) and are built up of one or more underlying actuarial layers that do the generic actuarial modelling. This removes ambiguity and allows the developer to appropriately group different features or characteristics.

 

What makes our assumption system different

Our assumption system works with a basis table as a driver of all other assumptions. The basis table receives a set of entries for each time period, with the entries being appended to the table and no old entries being replaced. The power in this is that all past valuation information is available at any point in time, and build-ups can be done easily without ad hoc manipulation of assumptions.

 

What the client said

“We were pleasantly surprised by the speed of transitioning, which was our concern when we started onboarding Insight and Mo.net. We were very impressed by the Insight team and the expertise presented. As we started working together, we found the team was very well organised, planned ahead of time for every task and had positive engagement.”

0

0 Comments

Get an email whenever we publish a new thought piece

Topic:

By signing up you consent to our terms and conditions

More from Christelle Oosthuizen

Email

Subscribe