Financial Model
Outcomes

MODEL APPLICATION RANGE

The largest data range that the model was validated was set at 20254, however the number of applicants may be as large as needed. When a dataset containing more than 10127 account numbers or new applications are received, queries would then need to be refreshed and formulae copied accordingly.

CLIENT OVERVIEW

The client who would be eligible for a payment holiday is one with a Nedbank loan for which they require said holiday. The client should not be overindebted, have a good credit record and the credit facilities that they have with Nedbank need to be up to date. Given that the cost of granting a payment holiday is high, as is the cost of default, any client with a history of default will not be eligible.

EXPLORATORY DATA ANALYSIS

Upon receipt of the data, it was cleaned whereby account numbers associated with clients younger than 18 were removed. As were those with empty fields, particularly the salary, instalment, credit score, start and end dates, and the number of defaults (if any). There were several fields associated with each account number, however the most pertinent variables for creating the model were gross monthly salary, monthly instalments, credit score, and default status.

MODEL FACTORS, WEIGHTS AND THRESHOLDS

The factors that the model considers in rendering an outcome are the debt-salary ratio, credit score and defaults. All the factors are weighted equally. A debt-salary ratio of more than 0.36 is considered as over indebtedness and these clients would not qualify. A credit score of less 640 disqualifies a client, as does a history of default.

COST TO BUSINESS

Cost to business using the old process (approving 50% of applications): R5 949 612,50

3 months payment holiday cost to business: R311 375

6 months payment holiday cost to business: R115 150

CONCLUSIONS/RECOMMENDATIONS

With 10 127 applicants, the model accepted 363; that is a 3,58% approval rate. While the model does drastically reduce the cost to the company by 92,8%, given that Nedbank is a people-driven company and seeks to help those clients whose accounts are in good standing, I would recommend relaxing some of the model factor thresholds. Of the three, the factor that I would recommending loosening is the credit score. Loosening the debt-salary ratio, i.e., making it higher, we run the risk of over indebting clients and increasing their likelihood defaulting. Since clients whose accounts are in good standing are the only ones eligible for the payment holiday, adjusting the default parameter requirement is not possible.