Credit scoring is widely understood to have immense potential to assist in the economic growth of the world economy. Additionally, it is a valuable tool for improving financial inclusion; credit access for individuals and micro, small, and medium enterprises; and efficiency. The use of credit scoring and the variety of scoring have increased significantly in recent years owing to better access to a wider variety of data, increased computing power, greater demand for improvements in efficiency, and economic growth.

The methods used for credit scoring have increased in sophistication in recent years. They have evolved from traditional statistical techniques to innovative methods such as artificial intelligence, including machine learning algorithms such as random forests, gradient boosting, and deep neural networks. In some cases, the adoption of innovative techniques has also broadened the range of data that may be considered relevant for credit scoring models and decisions.

The opportunities of using innovative methods for credit scoring include greater financial inclusion and access to credit, improvement in the accuracy of the underlying models, efficiency gains from the automation of processes, and potentially an improved customer experience. The use of innovative methods for credit scoring, however, also raises concerns about data privacy, fairness and potential for discrimination against minorities, interpretability of the models, and potential for unintended consequences because the models developed on historical data may learn and perpetuate historical bias. That said, there are also risks to consumers and businesses from a lack of innovation in credit scoring if it hinders improvements in financial inclusion and risk assessments. There are also concerns about the effectiveness of credit scoring methods and technologies. These concerns apply especially in markets with weak or no adequate regulatory oversight or industry codes to regulate the conduct of credit services providers (CSPs).

The guideline recognizes that the technologies supporting innovative credit scoring are still evolving and that differences in use, accuracy, and robustness exist across markets. The talent and data infrastructure required to execute the more innovative approaches are still very limited in many markets. The guideline encourages the adoption of a human-centric approach, where innovation is applied with the human in mind. This guideline proffers seven policy recommendations to guide on credit scoring, encompassing both models and decisions, in an effort to help regulators in their oversight roles and to aid in promoting transparency.

Credit Scoring Approaches Guidelines.pdf

Source: World Bank