Vose Software StopRisk

What is StopRisk?

StopRisk is a system of powerful mathematical and simulation software tools with which we can build for you the quantitative operational risk (OpRisk) analysis model that exactly matches your specific requirements. StopRisk has been used to build OpRisk models for banks and financial institutes to meet regulatory requirements in Australia, Germany, Nigeria, Singapore, the UK, and the USA.

StopRisk offers:

•Flexibility to model client’s unique requirements

•Completely configurable reporting of results

•Very wide range of built-in risk modeling and statistical tools

•Separation of risk modeling and management roles

•Can be linked to Pelican ERM software for consistency

•Can include archiving and access rights for audit trails

Modeling capabilities

The StopRisk system comprises a wide range of modeling capabilities:

•Over 30 different frequency distributions

•Over 60 different severity distributions

•Over twenty different univariate and multivariate stochastic time series models for risk factor modeling

•Fitting of distributions to data using MLE and automated fitting to maximize AIC, BIC or Log likelihood

•Ability to fit spliced severity distributions to data

•Ability to recalibrate to internal loss data any distributions fitted to external loss data

•Aggregation tools include Monte Carlo, Panjer, FFT, multivariate FFT and De Pril

•Ability to connect to loss records held in various types of database

•Correlation structures including Gauss, T, Clayton, Frank, Gumbel and empirical, all of which can be fitted to data using MLE

•Ability to include scenarios with loss distributions defined in various ways including tail shape factors

•Can be connected to our Pelican risk management tool, allowing monitoring and reporting of risk controls and mitigations and automatic updates when any controls/mitigations lapse or come into force

•Ability to apply multi-level business and/or regional correlation

•Ability to split by any combination of business entities, sub-entities, regions and sub-regions

•Ability to incorporate insurance effects

•Random generator seed control for replication

•Precision specification and monitoring

Commissioning process

Quotes of cost and delivery time are based on an initial scoping and participation exercise. Typical delivery times are 3-6 months. All models are provided with full documentation and training for the analysts and managers. All models are built individually to client specifications, after a detailed scoping exercise and determination of a testing protocol. Models are delivered in phases, typically:

•A review of the available cleaned data, the format and storage arrangement;

•A detailed scoping document describing the mathematics and algorithms to be used, the controls and features to be incorporated, the method of testing, the IT environment in which the model will be implemented, and any requirements for future-proofing;

•Delivery of a dummy model to be tested within the client’s IT environment. Solving any compatibility issues;

•Delivery of a draft model to be tested against known results and within the client’s IT environment;

•Incorporation of client’s comments within the agreed scope and correction of any errors;

•Delivery of a draft final model and documentation. Resolving any remaining issues;

•Delivery and installation of final model and documentation

•On-site or online training to modelers and managers

Highlight feature: Reporting

•Ability to define and run several scenarios and compare results statistically and graphically

•Interactive, formattable histogram, cumulative and sensitivity plots

•Report templates

•Numerical tables of calculations such as multiple VaR values, redistribution of VaR between business entities, etc.