Nordigen outperforms a major credit bureau in Spain
Our data team recently carried out a proof of concept for a client in Spain, with the task of helping the client evaluate the probability of their customer default using account information. In essence - turning account statements in to risk scores.
The results were stunning — not only did our scores prove to be very predictive, but we outperformed a major credit bureau by 12% measuring the GINI coefficient.
Here is an illustration of the results:
The clear advantage of account data is information about income and spending behaviours of private individuals. This information cannot be found in any traditional credit reports. On top of this, account data includes information about active liabilities — for example, monthly regular payments to a lending institution can be used as a proxy for an active loan. This is particularly useful in countries, where credit bureaus don't have a positive register.
That being said, there are two key takeaways from the test:
account data and credit bureau data capture different behaviours and the best results are achieved when both data sources are combined;
scores generated from account information can be more predictive than credit bureau data.
Have you tested account data in credit scoring models before?
Leave a comment below.
Nordigen helps banks and alternative lenders approve more creditworthy customers by leveraging account information — from helping lenders verify customer income to building complex risk scores from account information.
More information: https://nordigen.com/