Proposed Project for the MSU Industrial Math Students

Estimating the Workload and Credit Loss Impact of Collections Strategy Changes

Customers delinquent on their Ford Credit loans are handled by collections centers according to their risk of repossession. In other words, a customer who is unlikely to be repossessed will be allowed to go a significant number of days delinquent before receiving a collections call. By contrast, a customer who has a high likelihood of repossession will be handled more aggressively. The instrument used to assign a likelihood of repossession is called a behavior model. The behavior model is applied to a customer on every payment due date (one per month), and the model's prediction of risk is used to place the customer into one of many different risk groups, each with a different collections treatment, or "strategy".

Varying the time at which collections begin and the frequency with which customers are called will obviously affect the size of collections queues. As a result, the amount of work in the collections centers and the number of collectors required to do the work will be affected. Credit losses and delinquency rates should also change as a result of collections strategy changes. A simulation tool is needed to quantify the amount by which these very important measures will change when we make changes to collections strategies and risk groups. The simulation should be able to answer questions regarding delinquency, loss, workload, and required staff under any possible "what-if" scenario. It would be especially useful if developed so that it is friendly to "non-technical" users.

Top of Page