Retail Risk Modeling
Improve the portfolio performance by augmenting the retail risk architecture
Overview
Auronova Consulting provide full suite of all retail risk analytics solution to the clients. We also leverage external data available for increasing the discriminatory power of the scorecards such as Bureau information, other sources of data as may be applicable in context of the client.
Our retail risk analytics services are in compliance with regulatory requirements such as Basel II & IFRS 9.
Approach
A way of partitioning heterogeneous portfolio in a number of homogeneous segments, in response to differing preferences and past behavior then designing and implementing strategies to target them.
Segmentation can be used throughout the credit life cycle. It can be used as an alternative to scorecards or in tandem. Segmentation models group the clients’ applications/customers based on their profile, separating high risk customers from low risk customers, allowing appropriate decisions to be made. We help our clients in segmentation, which is typically an be effectively used in the following key activities.
- Application decisions (Accept/Reject/Refer)
- Score based decisions (different segments have different score cut-offs)
- Limit strategies
- Pricing strategies
- Limit increase/decrease strategies
- Cross sell/up sell strategies
- Collection strategies
- Marketing strategies
Application scorecards assign a score to each obligor such are for credit cards, mortgage loans, vehicle loans etc. based on the financials, demographics and some qualitative factors selected on the basis of the product for which scorecard is being made.
The score measures the risk associated with each new obligor by assigning a probability of default over a particular time horizon.
Methodology of Application Scorecards involves the following steps:
- Data collection and audit
- Time window selection and default identification
- Scorecard factor analysis
- Scorecard development
- Scorecard performance
- Scorecard validation
Application Scorecards provides a meaningful assessment of obligor characteristics and consistent quantitative estimation of risk.
It can be used in capital allocation, limit management, evaluating risk in portfolio, risk based pricing etc.
Behavioral scorecards enable timely monitoring of credit quality and improve predictive power.
It also assigns a score to each obligor based on payment history, financials and some qualitative factors, but measures risk associated with existing obligors whereas application scorecards calculate risk associated with new obligors. It mainly uses internally available information for risk assessment.
Behavioral scorecards are most predictive for smaller borrowers as they normally have only one banking relationship. They must also be developed for small businesses.
The methodology of this scorecard is same as that of Application scorecard with the difference of factors used to develop the scorecard. In addition to the uses mentioned for application scorecard, behavioral scorecards can be used for cross selling & up selling.
LGD is the loss incurred by a firm when an obligor defaults on a contract calculated as a percentage of exposure at default (EAD) that is unpaid post the workout period. It is calculated by discounting the cash flows including recoveries & costs, from the date of default till the end of workout period.
Types of LGD models:
- LGD Model for new Contracts
- LGD Model for existing non- defaulted contracts
- LGD Best Estimate Model for existing defaulted contracts
Model development methodology is same as that of scorecards with major difference in the development process.
It helps the firm by enhancing efficiency and effectiveness of recovery process and by performing meaningful differentiation of loss in case of default and consistent quantitative loss estimates. It can be used in estimation of regulatory capital for AIRB compliance, capital allocation, provision of Expected Loss, efficient resource allocation for collection effort.
Exposure at Default is the outstanding amount that obligor is yet to pay at the time of default for revolving facility like credit cards and line of credit products. EAD model estimates the outstanding balance of a contract if it defaults within next 12 months.
Types of EAD models:
- EAD Model for New Contracts
- EAD Model for low utilization contracts
- EAD Model for high utilization contracts
Model development methodology is same as LGD model. It helps the firm in measuring the future utilization of unused limit. It can be used in capital allocation and provision of expected losses.
Loss Forecasting Model estimates the future losses to ensure that current strategies are profitable and resilient to changes in the economic environment. Accurate forecasts of expected obligor charge-offs are essential for risk management. Separate Loss forecasting models will be required for each type of retail portfolio i.e. auto financing, personal loans, credit cards, home finance, small loans etc.
Uses:
- Make adequate provisions which is contingent on accurate forecasting of losses
- Provides better future estimates since it tracks movement of each contract
- Expected obligor charge-offs
- Provision of Expected Loss based on future growth
- Business Planning & Forecasting
- To enhance collection process efficiency & effectiveness
Key Business Benefits
Profitability
- Enables the bank to measure & better price the risk – ‘Improved Product Pricing’
- Profitability could be measured and compared consistently across products/segments
- Allows for selective ‘Up Sell/Cross Sell’ opportunities to the profitable clients
- Increase effectiveness of ‘Collection Strategies’ by using the LGD scorecards
- Allows for selectively increasing & decreasing the credit limit using EAD scorecard
- Increased discriminatory power of models will result in corresponding reduction in losses
- Portfolio risk segmentation allows for identifying the growth segments
Risk Differentiation
- Leads to increase in risk differentiation among obligors, products, segments, portfolios etc.
- Provides improved capability for identifying the clients for ‘Top Up and Refinance’ facilities
- Increase in the overall quality of the bank’s lending portfolio over time
- Enabling an automated credit decision making environment
Operational Benefits
- Improved ‘Turn Around Time (TAT)’ for credit approvals based on score cut-off strategy
- Credit function to focus only on ‘Borderline Credit Cases’ – Improved effectiveness/efficiency
- Proactive portfolio monitoring & collection management to limit the losses
- Enables advanced ‘Business Intelligence’ and ‘Analytical Insights’ to identify growth segments
- Helps to enhance ‘Risk Culture’ across the retail banking group
- Increased quality in data capture enables better data analysis and improved decision making