OpRisk modelling, management and capital requirements determination

OpRisk Science is a set of best of breed tools for OpRisk management through quantification and modelling which also permits the determination of capital requirements (Basell`s Pillar 2 and Internal Models under Solvency II). OpRisk Science  (OpCapital Analytics) have received 10 awards by Risk.Net and InsuranceERM including Best operational risk solution, Op risk scenario product of the year, Best stress testing producto and Best risk analytics tool (contact The Analytics Boutique for a demo and commercial details).

In addition, Chartis Research’s Enterprise GRC Solutions, 2019 classified The Analytics Boutique as:

  • Best of Enterprise Solutions in GRC Analytics Quadrant
  • Best of Breed on the Quadrant of OpRisk & Conduct Risk
  • Best of Enterprise Solutions in Model Risk Governance Quadrant

In addition, our methods have been widely published and our publications recognised by top practitioners (see Actuary Magazine and RiskBook).

Download The Analytics Boutique`s application brochures:

The Analytics Boutique differential value proposition in GRC is breaking the “bubble chart” paradigm by using strong quantification and analytics. By “bubble chart” paradigm we mean that GRC output should no longer be limited to nice dashboards of GRC basic metrics (i.e., loss collection, indicators, risk evaluations…). These GRC basic metrics are hard to understand by senior and upper management and too frequently end up ignored and not effectively integrated into daily management due to their unclear monetary value. Using quantification techniques (distribution models, forecasting models, correlations, Monte Carlo…), our solution transforms GRC basic metrics into GRC monetary value metrics (i.e., how much does risk cost, how much can be saved, NPV of risk mitigation investments, capital requirements…). Monetary metrics are the language senior and upper management understand and use daily. GRC metrics translated into monetary value can easily be integrated in the standard risk management processes: planning, budgeting, resource allocation and so on.

It can be used for the following purposes:

  • Collect scenario analysis
  • Determine the money value of Risk and the NPV of mitigation actions
  • Collect loss events and KRIs
  • Manage mitigation plans and build their the business case to justify required investments and insurance primiums
  • ICAAP and ORSA Operational Risk stable and robust capital estimates
  • Comply with ISO31000
  • Risk and control mapping
Structured Scenario Analysis
  • Great efficiency and strong model governance: workflow, approvals, complete audit trail, workshop invites, automatic reminders, answer aggregation, report generation and more
  • Cognitive bias mitigation: Group thinking, herding, authority bias, confirmation bias and others
  • Scenario planning and identification of most relevant scenarios
  • Priority scenario detailed development
  • Scenario support data in questionnaire including case studies, to support experts when answering
  • Possibility to answer form in workshop or, alternatively, answering each expert individually and aggregating individual answers into a single answer per scenario
  • Evaluation of risk scenarios under any number of dimensions: ie, financial, reputational and operative
  • Scenario scientific validation of qualitative estimates of risk using Structured Expert Judgment
  • On-the-fly Monte Carlo simulation for estimating loss profile by the first line of defence
  • Audit trail of origin and all transformations in the scenario down to the simulation report
  • Extensive model regulatory approval reporting
Action plans and mitigation analysis
  • Action plan management determining responsibles, reminders and alerts
  • Simulation of action plan NPV for building business case for required investments
  • Simulation of the impact of insurance policies permitting the detailed modelling of insurance features and calculating the NPV of insurance policies
  • Combination of insurance with action plans to determine the total mitigation in the risk profile and resulting NPV
Pre and post mitigation action loss distributions
Sensitivity analysis on scenario loss estimates
  • Automatically generates pre-defined sensitivity shocks for the scenario loss estimates
  • Permits scenario ad-hoc sensitivity shocks
  • Allows sensitivity analysis on distribution law assumptions
  • Automatically calculates sensitivity analysis capital results
  • Sensitivity analysis capital report
Global capital aggregation
  • In case of very large number of scenarios simulation, it is possible to aggregate scenarios using flexible aggregation paths in multiple steps
  • Each aggregation node permits the definition of a specific correlation matrix
  • Total capital is allocated down to all steps providing allocated capital to scenarios
Bayesian network modelling of scenario analysis
  • Most damaging scenarios or those in need of expensive mitigation plans can be modelled using our Bayesian networks user friendly and highly flexible capabilities
  • Results are integrated into the hybrid model together with the rest of scenarios or even LDA models
  • Bayesian network provides the pre-and post-mitigation analysis and the mitigation plan NPV, including insurance impact, in consistency with the rest of modelled scenarios
  • Bayesian networks approach is integrated with the rest of the workflow and governance framework: audit trail, permissions, aggregation and overwrite phases, reporting, etc.
Loss and Indicators Collection
  • Customizable loss and indicators collection forms
  • Unlimited number of collection processes for losses and any other indicators
  • Customised indicator (KRI, KPI, KCI...) collection workflows with reminders and alerts for faulty collection
  • Customization of indicator collection frequency and responsibles
Loss data analysis
  • Multiple incident management modules activated simultaneously
  • Multiple visual representation and analysis
  • Audit trail of all transformations, filters, etc., introduced in data
  • Audit trail of the data origin, filters and transformations which goes down to simulation report Frequency projections
Distribution fitting
  • 24 basic severity and frequency distributions
  • Mixtures of any combination of basic distributions
  • Simulations, fitting and comparison of all fitted distributions
  • User defined distributions via XML
  • Fitting via MLE, Robust Least Squares, Probability Weighted Least Squares, Probability Weighted and Moments approach
  • Distribution split in 3 segments (low losses, medium losses and tail losses)
  • GoF: AD, KS, Kramer von Mises, etc.
  • GoF visual: PP and QQ plots, histogram and CDF
  • Non parametric distributions
Extreme Value Theory
  • DEdH analysis
  • Tail parameter stability analysis by threshold
  • Tail plot
  • Mean Excess Plot
  • Hill estimator
  • HKKP-Hill
  • GoF analysis by threshold
Comprehensive hybrid model
A complete modelling of operational risk requires the accommodation of multiple loss types characterised by their loss frequency and criticality of the loss: high frequency, low frequency, worth modelling in detail using Bayesian networks or others that are not worth the detailed modelling effort. OpCapital Analytics hybrid model accepts all cases for a comprehensive and complete OpRisk modelling:
  • Internal Loss Data for high frequency risk categories
  • External Loss Data complementing tails
  • Scenario analysis based on direct loss estimates
  • Scenario analysis modelled in detailed using Bayesian networks, for those most critical scenarios
Monte Carlo simulation
  • Copula parameters fitting: Gaussian and t-student
  • OpRisk correlations calculation and stress testing
  • Nested copulas for the aggregation of multiple scenarios, data cells, etc.
  • Automatic stop when stability of results is reached
  • Audit trail report
  • Batch process for multiples consecutive runs with different parameterizations
  • Simulation in multiple currencies for multinational institutions
Insurance modelling
  • Modeling deductibles
  • Modeling maximum coverage
  • Applying deductibles and coverage for total losses or/and individual incidents
Operational loss forecasting, budgeting and stress testing
  • Time series loss forecasting models (ARIMA, ARIMAX and others)
  • Possibility to build operational loss forecasting models based on macroeconomic data or any other indicator (KRI, KPI, KCI...)
  • Projectionof operational losses under macroeconomic scenarios or under indicators scenarios, i.e., loss budgeting in case of attrition growing 20%
  • Strong times series model out of sample validation features
  • Modelling archive
  • Extensiveregulatory validation reporting
Backtesting of operational risk
  • QQ and PP plotscomparingnew losseswithdistributionsusedin capital estimates
  • GoF statisticscomparingnew losseswithdistributionsusedin capital estimates
Reporting in MS Office, PDF and RT
  • All graphics, tables and analytics have the corresponding report: Incident data analysis, Distribution fitting, EVT, etc.
Formal functionalities
  • Audit trail
  • User control
  • Interface with GRC software to import events, scenarios, user rights, etc.
  • Interface via ODBC drive
  • Handling of multiple currencies
  • Parallel computing for simulation
  • One click model replication
  • Modelling journal
  • Modelling archive


We always strive for excellence in what we do