Page d'accueil // SnT // News & E... // How Machine Learning Can Help Assess SME Credit Risk

How Machine Learning Can Help Assess SME Credit Risk

twitter linkedin facebook email this page
Publié le mardi 07 juillet 2020

A new partnership between SnT and Yoba Smart Money, a Luxembourgish technology start-up, will explore the use of machine learning to assess the credit risk associated with Small and Medium Enterprises (SMEs). Yoba Smart Money specialises in the research and development of risk management tools for the financial services sector using artificial intelligence.

SME’s Credit Challenges

The ease with which consumers can access credit does not exist for SMEs. Consumers can purchase almost anything on their credit card and easily obtain overdrafts, and they can do this because their payslip demonstrates their repayment capacity. For a typical consumer, once they have the paperwork, obtaining credit approval takes a matter of minutes. On the other side, the owners of SMEs live a very different reality that is frustrating and time-consuming.  Few suppliers accept credit cards, and lenders need to assess their credit risk by thoroughly analysing their business, which is expensive and takes time, so overdrafts are less common.

This is a problem as many SMEs have payment cycles that drain cash flow. Payment terms are often based on industry norms outside of their control or dictated by large market participants. When a business has a month with slow sales it can be forced to delay the payment of invoices to the next month. This can jeopardise important supplier relationships as well as potentially causing a domino effect further along the supply chain. This has been particularly problematic during the COVID-19 crisis.

 

Introducing Machine Learning

Yoba’s idea to solve this challenge for SMEs is to harness the wide availability of data today to support the assessment of credit risk for certain types of businesses. This is already possible in some countries, such as the Nordics, thanks to the availability of high-quality public data.  A credit risk rating that provides an assessment of the credit worthiness of SMEs is crucial to create a truly efficient credit market.

Yoba has therefore initiated a joint project with SnT to determine how traditional and alternative data sources could be used to perform automated credit analysis of SMEs in select European markets. Together with SnT researchers, Yoba will look at how technologies such as natural language processing, machine learning, and data modelling can be developed to be used for this purpose. The research project will determine the optimal data points that could be used to create an accurate credit rating, how that information about a client should be accessed and processed, and also create models that will allow the potential of an SME to default on a loan to be predicted.

Changing the Conversation

“One of the most frustrating things for entrepreneurs to experience is negative credit decisions, after waiting for that decision for months. Our proposed solution will assess customers automatically and in seconds,” said Atte Suominen, CEO of Yoba Smart Money. “The current way of working is not sustainable, and we can see companies, especially now during Covid-19 times, struggling with access to working capital financing. Despite initiatives to solve this challenge, such as Government guarantees that are meant to create easier access to working capital for SMEs, banks still have the same processes and assessment methods, which is the root cause.”

Atte Suominen, CEO of Yoba Smart Money

 

The Yoba Smart Money team has a strong background in both banking and entrepreneurship and has seen how the bottleneck in credit assessment impacts the economy. Thanks to their extensive experience and knowledge of credit granting for consumers they are well placed to help SMEs in this domain.

“We believe that we can tackle the problem of SME credit best through collaboration with the financial services industry and technology partners, which is why we initiated this project with SnT,” continued Suominen. “It is all about people with the right mindset and a professional way of working.”

“Working on industrial projects such as this pushes our research further as it gives us access to the real-world testing and data we need,” said Mats Brorsson, SnT’s principal investigator on the project. “Yoba’s proposed solution is compelling and could change the way SMEs are able to do business, which is exciting as it means our research will have a direct impact on the day-to-day lives of many people.”