The Credit Risk - Modeling ES bundle seamlessly and efficiently rates credit risks for global banks by creating highly accurate internal risk models.
These models rate default probability for bank clients, according to standards set by Basel II, and calculate the capital requirements required to cover high-risk assets. Because the models aren't standardized, risk managers can focus with laser-like clarity on each rating, thus more effectively allocating capital and reducing over-assessments.
The Credit Risk - Modeling ES bundle leverages enterprise SOA by providing integration with third-party modeling tools such as SPSS for Banking-Credit Risk Model Validation.
Risk managers at global banks who want to precisely rate default probability for their clients will find this ES bundle highly useful. The model pinpoints the exact capital requirements that are needed, based on Basel II international regulations, by building individual credit risk models and validating them to make sure they're accurate. Managers can then gain greater control over their risk rates when allocating capital. The result is increased profitability over time.
Credit Risk - Modeling (click to enlarge)
Previously, international banks used standardized models to calculate the capital requirements they needed to set aside for high-risk clients. Historical data was fed into a model with parameters set for risk-weighted groups rather than a single asset. That data could not be tailored, when toting up capital requirements for individual clients. So, the modeling process yielded results that ultimately overassessed capital requirements and produced inaccurate results.
The SAP Credit Risk - Modeling ES bundle offers integration with third-party modeling tools to help with creating models that assess individual risk rates for clients or groups of clients. This is done according to Basel II international banking regulations, where clients are judged according to rating classes. Next, a risk rate is computed. Once a rating is found for a client, the model is then validated over time to make sure it's still accurate. Because risk models can be calibrated again and again, managers can make sure that their models are constantly fine-tuned.
For details on Service Operations, Business Objects and Process Components, please check the ES Workplace.
Use Case 1: Building the Credit Risk Model
The bank decides to set up an internal credit risk model to predict the probability of default, the loss given default (that is, the loss if a default occurs), and the credit conversion factor. All of these elements are required by Basel II when creating internal credit risk models. The bank then adds any other features desired to its model.
Using a third-party modeling tool (such as SPSS for Banking-Credit Risk Model Validation), the first step is to build the model. The second step is to calculate the risk parameters for rating classes. For example, perhaps the best rating class, A, has a probability of default (PD) of 1% while the next class has a PD of 5%, and so on. Next, using a sample of data, you run the data through the model in the modeling tool, which invokes the Create Review Results enterprise service operation to store the results in the Historical Data Base (HDB) in SAP Bank Analyzer.
Use Case 2: Validating the Credit Risk Model
Having created the model, the next step is to validate it. Basel II requires periodic validation of models. You validate the model by running historical data through it. In this way you can compare the probability of default the model generates against actual default rates to see how accurate the model's predictions are.
The first step is to select historical data to use for validation purposes and then to run the model. Next you perform a statistical analysis of the model's predictions, comparing them to the actual results. For later analysis and to provide an audit trail, this data is then stored in HDB when the third-party modeling application triggers the enterprise service operation Create Review Results.
What if the model doesn't pass its validation, meaning that it isn't predicting well? In that case, the model must be calibrated, as described in Use Case 3.
Use Case 3: Calibrating the Credit Risk Model
Banks periodically validate their models, to make sure they are predicting as well as possible, and, then, if necessary, they adjust them, referred to as calibrating the model.
To calibrate the model, the users select historical data in SAP Bank Analyzer, then, in the modeling tool, apply the model. They then perform model validation with statistical analysis, as described in use case 2. If there is significant variation between the modeled values and the historical values, the next step is to adjust the parameters in the modeling tool. After rerunning the model, the third party application invokes the Create Review Results enterprise service operation to store the revised risk classes in SAP Bank Analyzer's HDB.
Future directions for this ES bundle will be market-driven.
- SAP Bank Analyzer
- SAP Banking Services 6.0
- A credit modeling application, such as SPSS for Banking-Credit Risk Model Validation