Solvency II valuations are generally complex, time consuming and potentially expensive The intricate calculations needed to meet regulatory requirements typically require a large number of data sources with varying quality of data, the involvement of multiple departments and teams, and a plethora of individual and manual tasks. Together, these can lead to very inefficient processes, with expensive actuaries often carrying out routine but laborious tasks when they could be much better deployed analysing results and adding value to the business.  At the same time, companies are also having to deliver more in shorter timescales, whilst reducing costs – putting a further premium on valuation efficiency.

Our software increasingly aims to remove inefficiencies and time constraints by automating all of the processes that don't directly require actuarial skills

Our software increasingly aims to remove those inefficiencies and time constraints by automating all of the processes that don’t directly require actuarial skills. This allows staff to focus purely on application of knowledge and areas of value add without worrying whether a process or task has been carried out when it should have.

Demonstrating this approach in practice, a recent Willis Towers Watson Solutions for Life proof of concept project has delivered the reality of going from raw data extracts to Solvency II quarterly reporting template (QRT) population in a single day.

Extending the envelope

In a very successful previous project for Chesnara (Chesnara Case Study), we had worked on streamlining the process of going from raw data to complete QRT reports. The results to date have been to reduce the working day timetable by over 65% and the resources required to produce the same results by 50%.

We set ourselves the challenge of going from raw data to delivery of the complete QRT reports in seven hours or less

But, in this case, we wanted to push the boundaries and demonstrate the levels of automation an insurance company could achieve with the correct application of tools and processes.  Therefore, we set ourselves the challenge of going from raw data to delivery of the complete QRT reports in seven hours or less, by automating processes and limiting the human interaction to when expert judgement was needed.

Tools used

End-to-end risk reporting process

End-to-end risk reporting process
Click to enlarge

The project used the following components of the Solutions for Life portfolio:

  • Unify: an enterprise risk and actuarial systems integration platform that enables integration with Willis Towers Watson software solutions and other third-party software. It provides automation of flexible, user-defined workflows to run these integrated solutions along with security and governance, process review and approval, version control and full audit trails.
  • RiskAgility FM: a financial modelling software product that enables life insurers to run financial models that accurately reflect their company’s products and to run them in ways that are easily adapted to their business processes.
  • DataValidator: a flexible and user-friendly software solution that allows a company to validate, cleanse and transform data (inputs and outputs) efficiently, preparing it for use in financial modelling and reporting processes.
  • vGrid: a software-as-a-service virtual grid solution offered in the cloud that provides RiskAgility FM clients with adaptable, on-demand compute power to execute financial model runs quickly and cost efficiently, helping them comply with regulatory demands such as Solvency II.

Process Automation

All the following processes were controlled entirely by Unify, which recognises what has already been completed and, based on that, can then determine what remains to be completed. It allows the process to run end to end in the most efficient way possible and to cope with almost any eventuality.

  • Data Process (Data Validator and Unify)
    • The data process took in raw policy data which equated to roughly 40 individual policy level files, in addition to an extra 20 files which included supplementary lookup data (notably unit prices).
    • These files were then validated and transformed into the required RiskAgility FM model input files by making use of Willis Towers Watson’s DataValidator software. DataValidator transformed and aggregated the raw policy data files through a number of complex rules, whilst also performing approximately 550 individual checks on the data.
    • These checks included validating the input data, verifying the generated model fields and performing movement checks on the data between valuation periods.
    • In addition to generating the required model files for RiskAgility FM, DataValidator produced data summaries for various reports, extra data-sets for processes such as the analysis of surplus, and specific information on certain policies required by the client.
  • Assumption management (Unify)
    • Unify provided Excel assumption files with instructions about what to change. This ensures consistency across the process.
    • Once updated, the assumptions were then returned to Unify ensuring full version control.
  • Liability cashflow production (RiskAgility FM, vGrid and Unify)
    • The liability cashflow process produced a full Solvency II base analysis of change and all the required stress cashflows runs across two funds (a ring fenced with-profit and a non-profit).
    • For the with-profit fund this required 40 stochastic projections using 1000 scenarios in each projection for approximately 35,000 policies.
    • For the non-profit fund this also required 40 deterministic projections across five sub-funds totalling approximately 275,000 individual policies.
    • This was made possible by using vGrid to provide highly flexible, on demand computational power across virtually unlimited cores – in this instance 1,360 were used.
  • Asset process (Data Validator and Unify)
    • The asset process used Data Validator to reformat, clean and check the multiple data files from multiple investment managers in multiple formats. The output from this part of the process is a single listing of data for each company, where the key fields required by the asset stressing process were populated appropriately.
    • The listing of asset data was produced by a spreadsheet parameterised to perform Solvency Capital Requirement (SCR) stresses for each line of asset data, provided the asset data is in a prescribed format.
    • For each market SCR stress, the last part of the process used Data Validator to summarise a large 100k+ listing of asset data into short summary tables , and to calculate the stressed value of unit prices that invest in those assets.
  • Deliverable production (Unify)
    • The required deliverables (both QRTs and management information) were automatically produced as results became available.
    • The reports are all in Excel spreadsheets and the process involves pushing the correct results into these spreadsheets and presenting for review.

Human interaction

The second success criteria was that actuaries could only interact in the process at points when expert judgement was required, for example when conducting reviews of results. All processes required multiple reviews in order to follow Professional Excellence standards (although the reviews were limited in nature for the purposes of this exercise). These reviews were then captured in the end to end audit trail produced by Unify. To replicate reality as closely as possible the actuaries were based in Manchester, London and Delhi. 

The outcome

The whole process was completed in 6 hours 37 minutes, proving that it is theoretically possible to run everything required for a significant sized Solvency II valuation within one working day

The whole process was completed in 6 hours 37 minutes, proving that it is theoretically possible to run everything required for a significant sized Solvency II valuation within one working day and across multiple time zones. The results exactly matched the real results from the previous valuation carried out for the client; the only points of interaction by actuaries were areas where they added value and all reviews were successfully stored in the audit trail.

Whilst this may have been a theoretical exercise and it was, to some extent, carried out in idealised conditions, it does prove that with thought and the desire to challenge existing practices and processes, married with effective software tools, step changes in performance and efficiency can be achieved. 

The bottom line for insurers could be significant time and cost savings, not to mention ensuring that technical teams can meet the ever more demanding requirements of both management and the regulators.