Buy-side Awards 2016
Modern insurance can feel like a cross between options trading and specialist lending. Many of the products created by insurers to meet customer needs, such as fixed indexed annuities (FIAs) and variable annuities (VAs), are complex to value and tricky to hedge. On the asset side, a low-yield environment is driving insurance companies to invest in a wider range of instruments, including less conventional products.
This is territory more familiar to banks. But those institutions – alongside their software providers – have spent years developing appropriately sophisticated valuation and risk measurement tools. Little surprise, then, that insurers with portfolios of complex assets and liabilities are looking to banks' tools in their asset and liability management (ALM) activities.
New York-based Numerix made its name as a developer of some of the most advanced asset-pricing and risk analytics for banks. Subsequently, through its engagement with insurers that sought its help in managing VA portfolios, such as Ohio National Financial Services and Transamerica, the firm has brought its modelling expertise to bear on complex liabilities as well.
The result is Oneview Insurance – an integrated platform for ALM and hedging that includes an economic scenario generator (ESG) for both risk-neutral and real-world scenarios – and the Leading Hedge risk management and hedging tool for life and annuity products. Oneview Insurance enables companies to apply consistent valuation methods to both liabilities and assets, including complex hedge assets, eliminating model risk between the two sides of the balance sheet.
"Until recently, there was very little stochastic modelling in the insurance industry," says Pawel Konieczny, vice-president of insurance solutions for Numerix's client solutions group.
"While mortality and longevity risk were well understood, the addition of guarantees and riders added on top of base life product contracts introduced new types of risk that were less familiar.
"Today, it's clear these guarantees and riders behave like derivatives, and have to be understood and managed like derivatives. This requires a core set of real-world and risk-neutral models within a stochastic modelling framework."
While seeking the expertise and tools of Numerix, insurers have also expressed the desire to incorporate their own views of potential macroeconomic developments into their scenario simulations. Numerix has responded by giving users the flexibility to customise the calibration of its ESG for their ALM and capital modelling.
"Some insurance companies have strong internal views they want to embed in their real-world scenarios – for example, having a set of stochastic rates hit target rates over a future time horizon. They may also have a volatility target around each point in time. These targets may be best-estimate targets or assumptions, or may be sourced or developed from other information, such as forward-looking economist reports and forecasts," says Konieczny.
Numerix has incorporated a two-factor Cox-Ingersoll-Ross model into its ESG for interest rate modelling and allows all parameters to be time-varying, including volatility-related parameters. "This enables users to achieve a desired distribution of interest rates that covers the entire projection horizon. Users are able to observe the distribution of real-world interest rates at each point of time in the future using a single calibration," says Konieczny.
New regulation is further driving the need for sophisticated ALM modelling wherever complex products are being offered to customers. Variable annuities are popular in South Korea, with many linked to the Korea Composite Stock Price Index, which has seen significant growth over the past several years resulting in increased premium revenues. However, regulation of guarantee reserving has changed since 2011, necessitating the use of a stochastic calculation approach with real-world modelling dynamics. This has recently led Seoul-based consultancy Risk & Actuary to partner with Numerix to incorporate flexible real-world modelling capabilities into the ALM and other risk services it offers.
We felt that partnering with Numerix was a better choice because it felt almost like an in-house development, while having the backing and stamp of a third-party company, as well as partially mitigating key person risk on our side
Jeroen Decuypere, Aegon
An aspect of Numerix's offering that a number of insurers and pension plan managers have found attractive is the flexibility of its technology architecture. In addition to the integrated Oneview Insurance platform, Numerix also offers the ability to integrate the platform's components with an insurer's in-house systems if they wish.
"Insurers often have production systems for their regular ALM calculations, but which are not good for ad hoc analysis. So they tend to do this analysis in spreadsheets, which are prone to errors and are cumbersome to handle," says Konieczny. Numerix offers an alternative approach whereby companies can have the flexibility for ad hoc analysis, but with the reliability and performance of a production system through the use of modern scripting languages, in particular Python and Matlab, to access Numerix's analytics.
Numerix has also been addressing the performance demands of modern ALM through the use of accelerated analytical techniques and high-performance computing, such as adjoint algorithmic differentiation (a technique for simultaneously calculating the sensitivities of a derivative along with its price) and graphics processing units (microprocessors designed for rapid parallel processing).
The performance and flexibility of technology and sophistication of models are usually key reasons clients cite for choosing Numerix's products for their ALM, alongside – critically – the company's willingness to help understand and solve clients' problems. This was the case with Aegon Group.
"We have chosen Numerix because of the wide variety of modules available out of the box, as well as the flexible and intuitive modular setup of the software," says Jeroen Decuypere, a Baltimore-based financial engineer at insurer Aegon.
"Also important were Numerix's clear engagement to tailor [the software] to our needs as much as possible and the fact its proposition did not appear as [much of a] black box compared with a number of competitors. The final choice was between creating an in-house solution or partnering with Numerix. We felt that partnering with Numerix was a better choice because it felt almost like an in-house development, while having the backing and stamp of a third-party company, as well as partially mitigating key person risk on our side."
The week on Risk.net, May 12-18, 2018Receive this by email