Clicky

X

Predictive Models in Consumer Lending: Loan servicing and resource management

The uncertainty regarding a borrower does not dissipate after the disbursement of the loan. For lending institutions who will also be servicing the loan, the servicing phase of the loan life cycle comes with a new set of potential risks and problems. For example, an institution may encounter operational or compliance problems that can affect the servicing of the loan which can carry a huge cost. Some of these operational problems arise as a result of poor resource management which can be fixed with the redesign of a firm’s training program. Predictive analytics can be used in the lending industry to help with loan servicing and resource management.

Predictive analytics for loan servicing

During the loan application process, lending institutions and underwriters review information (perhaps using a predictive model) to quantify the uncertainty the borrower presents. They evaluate the creditworthiness of the client to decide whether or not a loan should be disbursed. In most cases, the client’s probability of default is compared to the institutions risk threshold. If the perceived likelihood of default exceeds the risk threshold, then a loan is not extended. If it falls below the risk threshold, then the decision moves towards determining the appropriate loan amount, interest rate, and term.

Loan CycleIn the loan application and underwriting process described above, there is little consideration for the applicant’s payment behavior after the loan has been disbursed. It’s possible that an institution can evaluate a client, conclude the probability of default is close to 0, but the client fails to make payments on time, or worse the client defaults. Since the probability of default is the client’s credit-worthiness rating at the time of evaluation, it reveals little, if any, information about the client’s future payment behavior. Predictive analytics can be used to quantify the likelihood of adverse behavior by the client.

One of the most common adverse behavior by clients of the lending institutions involve late payments. Not only do late payments expose the institution into different types of risks like operational and liquidity risk, they can also be a huge cost for the institution. Predictive analytics can provide a lending institution insight on a client’s probability for late payments. The predictive model constructed to estimate this probability consists of the same borrower-specific financial attributes and macroeconomic indicators used to estimate a client’s probability of default. Information on a client’s future payment behavior is crucial as late payments, especially accumulated late payments, are the best indicator for a potential loan default.

Predictive analytics for resource management

A lending institution with insight on both the probability of default and the probability of late payments can make a more informed decision about the borrower’s risk. Of course, for a profitable institution, these probabilities will likely hover around the chosen risk threshold. This implies that an institution will inevitably disburse a loan for a client with a high probability of default and/or probability of late payments. Then, the question becomes, how do we use predictive analytics to improve loan servicing and reduce operational/liquidity risk?

The answer lies in the proactive segmentation of the institution’s workforce. By implementing non-arbitrary assignments to the institution’s loan specialists, servicing errors are minimized, while expertise and experience is maximized. For example, an institution can assign a group of specialists based on “tri-level” client risk rating. In this system, clients with an institution-defined high probability for late payments could be sourced to loan specialists specializing in high-risk clients. Similarly, medium-risk and low-risk clients could be sourced to appropriate loan specialists. This type of division of labor could not only help reduce servicing errors, but it could also allow the institution in developing a structured training process.

Predictive analytics when combined with creative thinking and sophisticated techniques can provide an institution insight that can lead to reduced costs and increased profits. In this case, a model can be repurposed and enhanced to create new meaning and derive new information from the same dataset. By quantifying the likelihood of late payments, the institution can evaluate the client’s profitability from an operational perspective. Using a combination of the probability of default and the probability of late payments, an institution has a glimpse of a client’s aggregate benefit and periodic benefit – two important pieces of information for planning and business management.

Tags:

Leave a reply

Your email address will not be published. Required fields are marked *