Risk Management Toolbox


Key Features

  • Binning Explorer app for developing credit scorecards
  • Credit risk simulation using copulas
  • Probability of default (PD) estimation using the Merton model
  • Concentration risk indices for identifying and controlling large exposure
  • Capital calculations using the ASRF model
  • Value-at-risk (VaR) and expected shortfall (ES) backtesting models for assessing market risk
Develop and scale risk management models, test against requirements, and run what-if scenarios.

Binning Explorer App for Developing Credit Scorecards

The Binning explorer app enables you to easily prepare data, apply automatic binning algorithms by just clicking, and manually merge or split bins. You can also save the credit scorecard that you are currently working on. You can then load the saved credit scorecard into Binning Explorer to continue working on the binning analysis. Alternatively, you can use the saved credit scorecard together with Financial Toolbox™ to perform other credit scorecard modeling processes such as: fitting a logistic regression model, formatting the credit scorecard points, scoring the data, calculating the probabilities of default for the data, and validating the quality of the credit scorecard model.

Binning Explorer app for developing credit scorecards, including the automatic binning algorithm, manual bin adjustment (merge and split), and bin information calculation.


Credit Scorecard Models for Consumer Credit Risk Simulation and Assessment

Consumer credit risk commonly arises from consumer credit products, such as a mortgage, unsecured personal loans, credit cards, and overdraft. By using Risk Management Toolbox together with Financial Toolbox and Statistics and Machine Learning Toolbox™, you can efficiently perform common credit analysis tasks, including

  • Perform interactive binning for credit scorecards using the Binning Explorer app
  • Create and analyze credit scorecards
  • Fit a linear model
  • Obtain points and score
  • Calculate the probability of default
  • Perform model validation

Validate credit scorecard models using the cumulative accuracy profile (top), the receiver operating characteristic (middle), and the Kolmogorov-Smirnov statistic (bottom). 


Credit Risk Simulation Using Copulas

Risk Management Toolbox provides a comprehensive suite of tools for simulating credit instruments based on copulas. You can simulate defaults only, or simulate credit rating migrations. Additionally, you can easily switch between Gaussian and t copulas. With these tools, you can determine the expected loss, value-at-risk (VaR), and conditional VaR (CVaR), or the expected shortfall. In terms of risk analysis and finding capital requirements, you can perform sensitivity analysis and stress testing on your credit portfolio.

Example histogram of simulated credit loses for a portfolio of credit instruments based on Guassian copula and t copula distributions.


Probability of Default (PD) Estimation Using Merton Models

One challenge in determining credit risk is in estimating core parameters like the probability of default (PD). Sometimes the risk manager derives PD based on the credit instruments' rating. Every company has its own unique capital structure, which in turn affects PDs for credit instruments. Risk Management Toolbox allows you to use Merton models to estimate PDs based on the capital structure of the firms you are measuring against.

The relationship between equity volatility and PD.


Concentration Risk Indices for Identifying and Controlling Large Exposures

It is hard to avoid concentration risk, as many businesses rely heavily on particular clients or segments. However, you may be able to choose to control or limit large exposures on particular segments. Using Risk Management Toolbox, you can calculate various concentration risk indices, including:

  • Concentration ratio
  • Deciles of the portfolio weights distribution
  • Gini coefficient
  • Herfindahl-Hirschman index
  • Hannah-Kay index
  • Hall-Tideman index
  • Theil entropy index

Example of Lorenz curve for representing the distribution of risk exposure.


Capital Calculations Using the ASRF Model

The ARSF function supports capital requirement and value-at-risk calculations using the Asymptotic Single Risk Factor (ASRF) model. You can calculate regulatory capital per counterparty using the analytical ASRF model by setting ASRF model parameters as described in the Basel documents.

Example of regulatory capital calculated by the ASRF model.


Value-at-Risk (VaR) and Expected Shortfall (ES) Backtesting Models for Assessing Market Risk

Risk managers often need to perform VaR and ES shortfall backtesting to assess the accuracy of their VaR and ES models. Risk managers are required to compare the level of profits and losses with the model-generated VaR and ES. The backtesting procedure is used to affirm the model validity. Usually, the risk manager uses more than one backtesting framework to perform VaR and ES backtesting on multiple levels of portfolios. Risk Management Toolbox enables you to perform multiple VaR and ES back tests.

Supported backtesting models include:

VaR backtesting:

ES backtesting:

  • Traffic light test
  • Conditional test
  • Binomial test
  • Unconditional test
  • Kupiec's tests
  • Quantile test
  • Christoffersen's tests

 

  • Haas' tests
 

Plot of the portfolio returns together with a 95% VaR from different VaR backtesting models to illustrate VaR violations (top), Example results from applying multiple VaR backtesting tests on multiple portfolios and VaR levels (bottom).