Matt Lightwood, director, quantitative finance, says that evolving the form and function of internal risk models is key to tackling the challenges in the market
With Solvency II now fully enshrined in law, what industry challenges have you sought to address in the last 12 months?
At Conning, we see at least three areas of focus in the European market. The first is insurers taking more ownership of the underlying internal model assumptions, traditionally a difficult task for actuarial departments to achieve. We have invested considerably in making the process easier and faster for the financial risk models embedded within our GEMS® Economic Scenario Generator (ESG), creating easy-to-use calibration tools backed by superfast computation of results. In many cases, the process of recalibrating models has been cut from hours to minutes.
The second is bringing internal modelling into the investment management process. Insurance asset managers need tools to more efficiently evaluate the appropriate asset allocation from a capital perspective and, as Conning is also an asset manager, the insight of our investment team has been valuable in developing these.
The last area of focus is automation and consolidation of process. We are just about to release a new automation workflow tool aimed at the life insurance market which will take much of the manual work out of the Market Consistent Embedded Value (MCEV) calculation and validation process. In testing, this is already proving hugely advantageous in reducing the end-to-end processing time.
How did Brexit, Trump and elections across Europe impact the underlying assumptions of your ESG?
These events led to many questions about appropriate changes for internal models. But since the actual events, things have changed very little. There is so much uncertainty about what Brexit means and what legislative success we can expect from a Trump administration that it would not be prudent to change our modelling assumption at this stage.
We believe that good risk models should not react to every political or economic event unless the event presents new, never-before-seen information. In fact, little in the way of new tail events has been observed so far post the Brexit referendum. We continue to monitor the situation and assess the potential impact, and I think so far this has been the right approach.
What is the legacy of Solvency II? Is it now business-as-usual for the industry or are people still grappling with it?
It is not business as usual. Regulators have approved internal models, and we now see them revisiting applications looking for evidence that the models are evolving and improving. As part of our response, we recently released a new interest rate model which addresses some of the difficulties life insurers have in fitting models to the swaption-implied volatility surfaces in stressed market conditions. I think regulators will really be looking for insurers to show that their internal models remain best of breed.
How much are low interest rates still dictating the agenda when it comes to risk modelling?
Interest rates still very much drive behaviour in markets on the investment side. Investors are still hunting high yield assets which in turn means they are having to accept more credit risk. We have expanded our models to build better credit risk analysis and are developing credit risk models for a wider array of asset classes. Our credit models now include European sovereign debt, corporate debt for multiple corporate sectors and US municipals, as well as several others.
New technology and buzzwords are now being introduced formally into insurance. How do you plan to develop in line with new tech?
Big data and machine learning are important concepts and we are looking at the benefits they offer, but we try not to jump on every buzzword and develop products around it. We have been working more on cloud and Software as a Service (SaaS) and will release a new cloud and SaaS suite by the end of the year, which should help to deploy technology more rapidly, increase performance and reduce cost.