Using unconventional data points to augment risk assessment for consumer lending (P2PL)

[avatar user=”Mark Petty” /]

Credit risk modelling is a pervasive and necessary tool used in lending to identity the likely credit worthiness of a Borrower.

Traditional credit agencies rely on historical data to predict the propensity to default over the length of the loan. The data points utilised for consumer credit risk have largely gone unchanged in the last few decades; defaults, credit history and transaction history such as phone, utilities and credit card repayments.

There are clear gaps in this current methodology. Timeliness of data one, but also a gap for the millions of consumers for whom credit agencies do not hold data on.

To have a truly fit-for-purpose risk model for P2P lending there is a need to draw on additional non-traditional datasets for proper assessment. Aggregation of social media, behavioural and big data analysis is key. The traditional credit agencies are talking about it, but like banks they are too slow to move and don’t have the expertise to innovate at the current time.

As we lead our lives online, the opportunities to augment traditional credit reporting with non-traditional sources increases. Everywhere we go we leave a footprint. Our connections on social media and the history of the products we buy can be combined to construct a picture of who we are and our likelihood to payback a loan.

ZestFinance and Kreditech are a part of the new wave of credit agency (and further proof that financial services is being disrupted from all angles). Both utilise big data analytical techniques to mine social media networks and ecommerce transactions.  Kreditech claim to analyse 15,000 individual dynamic data points, contrast that with traditional credit bureaus who have at most 10 to 20 variables.


Leave a Reply

Your email address will not be published.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