In most Sub-Saharan Africa lending to retail customers has not seen the growth rates witnessed in other regions in the world. The main reasons for this lack of credit offers on the supply side are related to high operational costs and to the general lack of credit history for large segments of the population. The traditional credit underwriting processes and data sources cannot help to change this situation.
About our Client
Our client (Commercial Bank – CB) is the market leader in its home market, an East-African country. Less than 30% of the adult population is formally banked effectively limiting the access to credit for large segments of the population. However, also the banked segments are not always receiving access to credit products due to high operational costs and a lack of data required for underwriting.
CB serves more than 500,000 customers and is leading lender for most segments. In order to defend its position in the market against competition from new entrants, it decided to tap into the potential of the unbanked and underbanked. This change in strategy required the bank to develop new credit products and to introduce new technologies and underwriting processes. To achieve these goals, the focus had to be put on solutions that provide both – lower operational costs and use alternative data sources to bridge the information gap.
CB decided to develop a mobile loan product which would be made available to external customers using a mobile wallet as well as to its internal customers. By entering into new partnerships with the leading mobile phone operators, it created a potential market size of more than 1.2 million new customers for credit products.
Paretix was contracted to provide the analytical components required for the new credit products. The supplied Paretix Smart Lending solution consisted of algorithms to score customers, a platform to handle the application flow and interfaces to the different data sources (internal and external).
The analytical challenge consisted of scoring customers based on their history of transactions. Most financial institutions (as well as mobile operators) need detailed transactional data for their day-to-day internal operations. While accessing this data is easy, the efforts to apply analytics are high. To calculate a meaningful credit score, the behavior of customers has to be tracked over a period of 3-6 months in the data. This includes the analysis of each transaction made either on the mobile wallet or in the banking account. For some customers this can add up to more than 1,000 transactions per month which have to be evaluated in real-time.
Together with CB’s credit experts, we developed high credit underwriting models. Due to regulatory limitations, the innovation had to be introduced by how data is prepared and used for the underwriting purposes rather than by introducing better modeling methodologies. Paretix’ experience in dealing with high-volume transactional data (e.g. from credit card companies) was what the bank needed: In a short time-frame we calibrated our scoring models to the different types of transactional data used for credit underwriting. In total, more than 50% of weights are given to transactional data.
The performance results convinced CB to move forward with the project and integrate the solution.
The implementation of the scoring algorithms handling high volume data can be challenging. The Paretix Smart Lending Platform provides the ideal environment for hosting such procedures and at the same time provides additional features like integrated back office, product configuration module and flexible ETL tools to connect to the different data sources. The built-in machine-learning capabilities ensure that performance is kept over time with low maintenance requirements.
CB launched the new products and attracted during the first months more than 30,000 customers new to credit. The average loan size is more than USD 130 with loan durations ranging from one to 12 months. The growth rates are in line with business expectations.
Due to the implementation of a truly digital lending process, the customer receives an instant credit offer available for instant disbursement. This advance in technological abilities helped the bank to improve its image and to regarded as a financial innovator in the region.