cFrame® is a flexible platform for modeling insurance products. It allows a company to develop realistic models to suit specific needs, and it serves as a development platform for a company’s own model that generally relates to analyzing stochastic characteristics of individual contracts and portfolios. As a generic simulation engine, cFrame is suited for insurance (life, pension, health, or non-life), and for single company or group modeling.
cFrame users can define insurance contracts as a set of cash flows that depend on user-defined random variables such as death, additional payments, or retirement. All definitions are done at a high level and require no low-level knowledge of code vectorization and parallelization. cFrame provides an interface for managing all model definition modules and data sources. Contract simulation models can be integrated into an existing economic scenario generator to model contracts that depend on financial markets or the economy at large.
The platform was developed from the start to support all regulatory (e.g., Solvency II) requirements and it can be used both for implementing the standard model and developing one's own internal model. Such a model allows market consistent embedded value (MCEV) calculation and solvency capital requirement (SCR) calculation. Further, the same model can be used for real-world risk management including own risk and solvency analysis (ORSA).
The whole model structure can be created in the interactive graphical interface and stored into the internal cFrame database. This allows both user friendly model building and easy collaboration between actuaries. In addition, all model definitions, user functions, and input and output data are stored into the same database. This method creates an unbreakable audit trail and allows exact replication of previous simulation results.
cFrame is a MATLAB® toolbox that supports the parallel computing capabilities in MATLAB but that can be used in a single-core environment. cFrame requires no additional toolboxes, although Database Toolbox™ may be needed to feed the data for the model. It is possible to speed up the model with a single multicore (up to 8 cores) workstation with Parallel Computing Toolbox™ or by using a computing cluster with MATLAB workers.