The argument for using a cloud platform for capital modelling

By: James Maudslay - 02/04/2012

James is responsible for developing Colt’s services for the insurance sector. He has worked in the London Insurance Market since 1989 and is a fellow of the Chartered Insurance Institute.

  • Like this
  • Close

At Colt, we are working with a number of customers in the insurance industry to support more dynamic decision making for underwriters. Cloud services can improve access, processing and delivery of the information essential for the business.

Wherever information is found in a modern underwriting firm, there will usually be an actuarial team on site to provide analysis services to the business users. Actuaries obviously require high spec PCs because of the data crunching that is essential to their roles but when the Board needs to decide whether to underwrite a new line of business or not, the last thing they want to hear is that the model doesn’t run fast enough, or it has fallen over because the bank of computers in use don’t provide a stable platform. The business needs quick and reliable access to management information that allows it to deliver risk adjusted return on capital to its shareholders. Equally, if a capital model takes 3 or even 6 hours to run then the actuary needs to be free to carry on with the rest of the job while the model runs, rather than being impacted by reduced processing power. 

In the last 6 months, we have conducted cloud computing trials for both capital and catastrophe modelling. The response from participants, both risk analysts and actuaries, has been overwhelmingly positive. For example, by working together with actuaries from a number of different underwriters, we have been able to design and build platforms that provide the stability to complete the huge number crunching exercises required for capital modelling. A stable platform that can scale as required provides confidence that information can be delivered to the business within a requested timescale. 

While confidence in on-going stability is key, the issue of speed is also vital for capital modelling.  When the opportunity arises to underwrite a new line of business, data obviously needs to be made available to understand the amount of capital required and whether changes to the existing quantity or make-up is required. Waiting a weekend for a model to run before the decision is made leaves the business open to the competition gaining first mover advantage.

One of our customers had a specific requirement to reduce the model runtime to a maximum of 3 hours. Working together, we set up a pilot to see whether the application could perform successfully within this window. Once the environment was up and running, the first phase proved that it was a lot more stable than the previous approach of networked PCs. The next step proved that it would meet the business objectives of running the programme in less than 3 hours, again and again. The team is now able to run the models twice in a single day, allowing the business to make decisions more quickly and respond to market opportunities. Considering that the 3 hours includes data transport across the internet, we have worked out that if we connect the office and the cloud using a dedicated link, we can reduce the running time to just over an hour and a half.

As the complexity of datasets grows, the processing power needs to expand, which could mean throwing more high spec servers at the problem every 9 to12 months. Given procurement processes and usage levels, neither the actuary or the IT manager want to be confronted with these constant additional requirements. As well as guaranteeing confidence that data can be provided as and when the business requires it, a cloud platform also provides infrastructure that meets the continuously growing processing requirements of capital modelling without the upfront investment.  A cloud platform offers a financial model based purely on consumption, a pay as you go approach that absolutely works for models that need to be run on a monthly or ad hoc basis. Solvency II will almost certainly require capital models to be run on a more regular basis, and internal business planning may dictate a monthly requirement or an ad hoc service throughout the year.  Whatever way you look at this, regular periods of non-use occur, and paying for a dedicated high spec environment 365 days of the year doesn’t stack up.

Our customers are running these models to make business critical decisions while at the same time trying to evolve their business processes to address Solvency II requirements.  In an environment where insurers are increasingly analysing high exposure risks, they need a platform that gives them the ability to be agile in their decision making.  Stability, scalability and pay as you go are all really important elements of why cloud services can support the business.

From our perspective, cloud services are not a new technology. They present a different way of consuming IT services that is more in line with the flexibility businesses today need to address their challenges. It changes the role of technology within the business, making it an enabler for the broker to get to market faster or improve its internal processes to be more effective.  As the insurance sector continues to grapple with the increased capital and reporting demands of Insolvency II, these trials prove adopting cloud computing platforms significantly reduces the challenges of running capital models.

This article will be available to read in the 5th April 2012 print edition of The Actuary.

No comments on 'The argument for using a cloud platform for capital modelling' Be the first to comment

Add your comment

Terms and Conditions Submit

Please choose your country