When there is a structural break, traditional credit risk models lose a big part of their predictive power. Applying transactional data and machine learning techniques, on the contrary, may result in much more reactive models without losing any efficiency in forecasting probabilities of default (PD), provided such models guarantee interpretability and explainability, Giangiacomo Sanna, Senior Specialist, Prometeia, and Marco Stella, Partner, Prometeia, found.
The Covid-19 pandemic has been for banks’ credit departments a textbook case game-changer, and we still haven’t seen all of its effects. It accelerated credit risk modeling for banks that are willing to know which company to support first in times where moratoria and public guarantees may hide the real current situation of the borrower and postpone the showdown with NPLs.
This message was at the foundation of the last Prometeia virtual session about ‘Improving PD models’ predictivity and reactivity with transactional information and machine learning’, where we presented our recent project developed with UniCredit – one of the biggest European banking institutions – to create a PD model for the small business segment. This process, which leveraged more than 550 million transactions, used advanced text analytics techniques to categorise free-text transaction purposes, and machine learning to turn categorised transactions into a score for each firm.
This cutting-edge model is proving to be well performing during the Covid-19 pandemic, even stand-alone, and it is very reactive, able to capture deterioration signals right from the start of lockdowns and restrictive measures. With a strong real-time component, it is able to differentiate transactional credit scores from more traditional ones, that is allowing banks to detect companies that need the bank’s support more effectively, even without a visible credit deterioration.
The use case presentation, by Emanuele Giovannini (UniCredit) and Giangiacomo Sanna (Prometeia), preceded a high-level discussion among some very qualified matter experts in the field: Chiara Capelli, Head of Credit Risk Modeling at UniCredit, Dmitri Kraynov, Head of Risk Modeling at Sberbank, Sid Dash, Research Director at Chartis Research, and Marco Stella, Prometeia’s partner.
Among other issues, the reactivity of transactional and AI-based credit risk models was cited as one of the most crucial features nowadays, especially in times of Covid-19 and accelerated digitalisation. The increase in predictive power is a big advantage, too, particularly for clients who usually leave little ‘official’ real-time information about themselves (e.g. small businesses, individuals).
Not surprisingly, the use of transactional data in risk modeling has lately been strongly recommended also by the ECB, the Eurozone regulator about banking risk models.
Such an implementation implies several challenges, of course, from an ethical to a technological standpoint. For instance, organising such broad sets of data requires a great computational power. But probably – this has been one of the loudest messages of the panel – the main challenges are around the explainability and interpretability of models.
Beyond regulatory and business needs, these features have proved to be a modeler’s exigence as well. During the pandemic, adapting the credit rating model has been only possible because of its interpretability, which has allowed to look inside the model and exclude the variables that were not reliable any more (for example the payment of taxes in Italy, where they have been postponed by authorities, and so lost their predictive meaning).
Moreover, interpretability has been said to be crucial to change the attitude of many internal and external stakeholders of the bank, who are extremely diffident towards black boxes. This stance makes investing in explainability even more compelling.
One thing is for sure. The road traced by these complex techniques can open up new business horizons for intermediaries. Think, for example, of the extension from transactional data to PSD2 and open banking data. A wealth of information that, framed in a model that can be explained and interpreted at every single point, could mark the new frontier for advanced loan origination and monitoring processes.
To watch the webinar again visit here.
This webinar is part of the Prometeia AI 4 Risk Expert Circle initiative, the series of digital roundtables thought for the community of risk practitioners, dealing with the new frontiers of Artificial Intelligence, Machine Learning and broader Data Science techniques applied to Credit, Financial and Operational Risk. Check out Prometeia’s website for further details to join the circle or write to: email@example.com.