Risk Management

Financial Risk is defined as the uncertainty of a capital investment. Our systems for risk management aims at estimating and managing the main financial risk types of a financial branch. Risk distribution and corresponding risk capital allocation are assumed to be as follows:

  • Market Risk - 15%
  • Credit Risk - 50%
  • Operational Risk - 35%

Other risk types, in regard to the last economic crisis, include liquidity risks of assets and spread risks of assets and derivates. Risk Engine and Risk Framework cover a set of internal models that calculate risks measured and managed financial institutions.

Architecture of our solution for risk management

Our solutions for risk management and risk mitigation use technological, human and organizational resources and follow international principles and standards (ISO 31000). Risk management includes identification, assessment, and application of measures, in order to minimize, monitor, and control the probability and/or impact of loss events.

Our Risk Management solution covers the most relevant risks, both in the financial and in the non-financial branch, since risks in non-financial institutions can be expressed in the last stages, in terms of financial risks. For example, the risk of price changes in materials, buildings, services, etc., is in fact market risk, in the same way as risk of changes in share prices.
The architecture of our solution shows a straight structure for calculation of various risk, integrated with other supporting modules within the entire risk management system. The main modules for Value at Risk (VaR) calculation include:

  • Calculation of market VaR via Monte Carlo simulation
  • Calculation of credit VaR via Monte Carlo simulation
  • Calculation of operational VaR via Monte Carlo simulation

A set of supporting modules is used to produce the needed calculation data from primary market data:

  • Rating tool box to produce ratings
  • PD and matrix calculator to calculate the rating based default and migration probabilities
  • Portfolio management to construct portfolios and calculate portfolio positions
  • Loss Given Default (LGD) simulator to produce data in case of default of an exposure
  • Operation risk module to simulate operational risk losses
  • Market risk module to simulate and aggregate market risk along an instrument specific list of market factors
  • Reporting modules that produce OLAP & Standard reports for screen, printer and export files
  • Settings models to configure scoring and rating models
  • Models for validations (GINI Diagram) and analyses of scoring and rating models using neuronal networks

Modules Portfolio management (Link to Portfolio Management in Risk Engine) and Market risk, via Monte Carlo and/or historical simulation (Link to Market Risk in Risk Engine), are presented in Risk Engine as well.

Risk Evaluation Data Flow

The data flow shown ensures the calculation and simulation of various risks, as well as the preparation of primary data, coming from the evaluation environment outside the system:

  • Rating levels are calculated for every issuer, based on the balance sheet data (hard facts) and a set of parameters and classifiers (soft facts). Rating levels can be imported from external sources, as well. Produced ratings are assigned to issuers and are used in the subsequent risk evaluation.
  • The used scoring and rating models are validated and analysed for reduction of the relevant factors and optimized rating structure by neuronal networks
  • Probabilities of default (PDs), migration probabilities (MPs) and migration matrix are produced based on historical time series of rating migrations or from historical statistics of cumulative PDs. PDs, MPs and migration matrixes can also be imported from external agencies. PDs and the MPs are used to simulate credit risk loss distribution.
  • The portfolio management module displays the risk evaluation of a portfolio subject and rolls out its cash flows. Exposure at default (EAD) is calculated based on market data.
  • The Loss Given Default (LGD) simulator models the specific LGD exposure, based on historical loss data, as well as on collaterals assigned to the exposure. LGD is then used in the credit risk evaluation process, in the default case, to account for recovery and for risk absorption by the collaterals.
  • Module operation risk receives Key Factor distributions from the Monte Carlo simulation and simulates operational risk losses within an institution’s hierarchical structure. Key Factor distributions are mapped to theoretical distributions, using primary information from the historical loss data base and the self-assessment module.
  • Module market risk operates on fast simulation trees of portfolio positions. Market factors for the Monte Carlo simulation are obtained from historical time series and their statistic features are held in market risk correlation matrixes and volatility vectors. The module simulates and aggregates market risk along instrument specific lists of market factors, on position and portfolio level.
  • All risk results and figures are stored in common report database tables, which are then reported via reporting modules, producing OLAP and Standard reports and exports.


The functionality of regulatory requirements is ensured via the following modules and models:

Module List Models and Functionality
  • Risk analysis - internal models
  • Market risk:
  • RiskMetrics model
  • Monte Carlo Simulation
  • Historic time series and volatility of market risk factors
  • Correlation matrixes and Volatility
  • Calculation and aggregation of Multi-factor Monte-Carlo
  • based VaR types:
  • Total, FX, Interest rate, Stock Index, Share, Mixed,
  • Marginal, Incremental, Spread, Historic
  • Market VaR projection over time
  • VaR break down by asset classes
  • VaR back testing
  • Multiple market scenarios applicable
  • Risk analysis - internal models
  • Credit risk:
  • CreditMetrics model
  • Monte Carlo Simulation
  • Credit Risk+ Model
  • Credit risk factors:
  • Probability of default (PD), migration matrixes
  • Seniority classes, market spreads
  • Industry sector time series and correlation
  • Calculation and aggregation of position VaR, issuer VaR,
  • expected loss, marginal VaR, incremental VaR
  • scenarios can be used
  • Risk analysis - internal models:
  • Operational Risk:
  • Advanced Measurement Approach (AMA)
  • Monte Carlo Simulation
  • Advanced Measurement Approach (AMA)
  • Internal operational risk data model
  • Stochastic models for severity and frequency
  • Correlation of key risk indicators
  • Loss distribution simulation
  • Business Line (8 lines) allocation according to Basel III:
  • Investment Banking
  • Corporate Finance, Trading & Sales
  • Banking
  • Retail & Commercial Banking, Payment & Settlement
  • Others
  • Retail Brokerage, Agency Services & Custody
  • Asset Management
  • Calculation of loss distributions from self-assessment and loss events
4 Rating and scoring
  • Based on rating and scoring models:
  • Corporate rating, Private rating, Bank rating
  • Balance sheet of corporate customers
  • Private, corporate and collateral scoring, credit allowance
  • Solvency II: Insurance risk
  • Market risk
  • Default risk
  • Life Underwriting risk
  • Insurance Instruments, mortality and longevity tables
  • Regulatory settings, market data and solvency attributes
  • Calculation structure:
  • Market risk, Default risk, Life Underwriting risk
  • Solvency II Capital Requirements, scenarios can be applied
Solvency II Coverage

The current coverage of Solvency II functionality is shown in the figure below. The coverage includes market risk, default risk, life underwriting risk and the calculation of BSCR (Basic Solvency Capital Requirements) and BCR (Basic Solvency Capital Requirements). Тhe calculated Basic Solvency Capital Requirement (BSCR) is adjusted for loss absorbency, using Equivalent Scenario or Modular Approach for each risk category. Solvency Capital for Operational Risk is considered in the calculation of Overall Solvency Capital Requirement (SCR) and Minimum Solvency Capital Requirement (MCR).

Rating and scoring evaluation

Ratings and scoring, within one or several internal or external rating systems, estimate the credit standing of counterparties. Rating estimations are needed as a measure for the ability to serve loans and other debt transactions in the credit risk calculation. Risk Framework offers solutions for estimation of private and corporate ratings, based on balance sheet data and soft factors. Country specific factors and criteria can be defined to consider particular economic environment conditions. Scoring results can be mapped to rating levels, using unified master rating scales. Certain rating levels are determined by early warning signals and corresponding KO-criteria. Official external information about counterparties can be included. In addition, scoring models evaluate the quality of the applied collateral, as the relation between planed debt redemption and debtor incomes. In this way, scoring can be used within the loan allowance procedure.
The quality of the produced rating and scoring is enhanced by adjusting rating and scoring models based on historical losses of a debtor pool, as well as produced ratings and PDs of the rating and scoring modules.

Advanced Measurement Approach (AMA) for operational risk

Advanced Measurement Approach (AMA) for operational risk is calculated via the Monte Carlo Simulation, using the Copula approach on non-normal distributions. Simulation distributions for severity and frequency for every loss event type are adjusted, based on the economic capital. This capital is calculated either according to those loss event types or to historical loss data base that accumulates internal or external loss data for the past 5 years.

Future Developments and Extensions

We permanently adjust and improve the implemented financial and mathematical models for capital risk evaluation and control, following new standards and regulations. Additional models are currently being developed, for example:

  • Operational Risk: Calculation of loss distributions from self-assessment and loss events.
  • Introduction of the Importance Sampling Approach in all simulation modules of Risk Engine and Risk Framework, thereby dramatically increasing the precision of the Monte Carlo Simulation (over 100 times).

Interfaces and Connectors

Module “Risk Management” is part of the entire risk management system. The module inherits the features of Risk Framework Interfaces and Connectors and Risk Engine Interfaces and Connectors.


1. Which market data is needed for the calculation of instruments?

Actual data for every interested currency:

  • FX rates to leading currency, e.g. EUR/USD or GBP/USD.
  • Market curves for usual market segments, such as UDS-Money, USD-Swap and USD-Capital.
  • CDS curves and CDS indexes (CDX) of CDS premiums for money and capital market.
  • Market indices of stock or industry markets, such as FTSE or DAX.
  • Market prices of securities, such as shares, funds, bonds, etc.

Additionally, for pricing of options or embedded options, implied volatility data is required. A set of relevant market factors is obtained from the pricing model for every separate instrument type.

2. Which market data is needed for the Monte Carlo simulation of market VaR?
The simulation is based on a correlation matrix and the volatility of market factors, so for the market data given in FAQ 1, daily time series for the last year have to be provided.
3. Which market data is needed for the simulation of credit VaR?

The market data given in FAQ 1, in addition to credit risk:

  • Migration matrix of rating agencies that contains rating movements or default probabilities, e.g. Standard & Poor, Moody’s or Fitch.
  • Collateral data and recovery rate classes used to calculate Loss Given Default (LGD).
  • Industry indexes used to calculate the industry index correlation matrix.
  • Yield spread curves for rating levels.

For every issuer or contract, the following assignments are created:

  • A rating level within a rating agency, e.g. BB rating within the Moody’s matrix.
  • Collaterals and recovery rate class.
  • Dependency of one or more industry indexes, given as percentages.
4. Which internal simulation models are used for market, credit and operational risk?
  • For market risk: RiskMetrics using Monte Carlo Simulation.
  • For credit risk: CreditMetrics using Monte Carlo Simulation.
  • For operational risk: Advanced Measurement Approach using Monte Carlo Simulation.

All models are based on the Gaussian Copula, so they can work with non-normal distributions.