PayNet puts its vast database and scoring expertise to work by providing state of the art scoring services to maximize business potential and minimize risk. From Score Performance Analysis & Scoring Strategies to Score Monitoring and Custom Score Development services, PayNet is a comprehensive partner, and component, to the Risk process.
Score Performance Analyses are essential to evaluate how well a score fits your business. For lenders who are already members of PayNet, this analysis may be completed based on the information already provided as part of the monthly reporting process. For lenders that are not members of PayNet, a file containing originations over a given period of time is required.
Often referred to as a Retro Analysis, this powerful analysis compares the distribution of scores at origination with the corresponding defaults of the population in order to compute the predictive lift that a score provides. An output of this analysis, the Lorenz Curve, allows a lender to better understand how the implementation of a score will affect their business either through more profitable (i.e. less risky) transaction identification or via increased originations while holding the level of risk consistent.
So, the best score for a lender's business has been selected. Now, what is the best way to implement that score into the business? Scoring Strategy analysis, often referred to as Swap Sets, can go hand in hand with a Retro Analysis. A Swap Set translates a Lender's Risk Strategy into actionable economic terms.
Credit application data is required to complete this analysis including customer name and demographic attributes as well as the credit decision and the booking status of the application. Patterns may emerge as the analysis is completed that will drive different approaches to particular segments of the business. These patterns give rise to insights on transactions that may not perform as expected and missed opportunities. Due to the depth of the PayNet database, the possibility exists to identify transactions declined by one lender but approved with another, or missed opportunities, and their ensuing performance. All of this information facilitates a targeted review of how the score is applied to specific transactions thus driving the Risk Strategy of the business.
Score Monitoring is a critical process employed by any lender that utilizes credit scoring during the underwriting process and should be completed on a regular basis.
The Score Monitoring process evaluates how well a lender's score, at time of origination, predicted an event of default after a prescribed amount of time has passed. This information can then be applied to the business strategy to adjust the score cut-offs used in the underwriting process or to evaluate how well a score fits their business.
Often times, lenders will implement a score – either a custom score or generic model – which they have found to be applicable to their business with the expectation of better managing risk on a transactional, and by extension, a portfolio basis. While scores effectively rank order borrower risk, the associated default rate for that score may shift over time due to multiple factors. Therefore, while a lender believes they have been consistent with their Risk policy through an economic cycle, they may have actually allowed more risk into their portfolio than their comfort level. Score monitoring helps to identify the shifts in default patterns thus providing the lender with better information on which to make policy decisions.
PayNet's extensive score building expertise is employed to derive the best scoring solution for each lender. Guidance is provided through all steps of the score building and implementation process from the initial stages regarding what attributes are the most predictive through the implementation of the score within a lender's business.
The process helps lenders evaluate various data sources in an effort to find the best fitting external data for their business. In some cases, a single data source may not be sufficient and will have to be paired with other information to provide the most predictive lift from the model.
At the end of the process, a lender will have an Empirically Derived Statistically Sound (EDSS) score model to help manage risk within the organization. But the process does not stop here. Assistance with developing appropriate Business Rules is also provided which includes things such as score cut-off setting, advance rate, term, etc. The Business Rules can help govern the application flow focusing the Risk resources to the transactions that need additional analysis before a final decision can be made.