Skip to main content

2022 | Buch

Artificial Intelligence and Credit Risk

The Use of Alternative Data and Methods in Internal Credit Rating


Über dieses Buch

This book focuses on the alternative techniques and data leveraged for credit risk, describing and analysing the array of methodological approaches for the usage of techniques and/or alternative data for regulatory and managerial rating models. During the last decade the increase in computational capacity, the consolidation of new methodologies to elaborate data and the availability of new information related to individuals and organizations, aided by the widespread usage of internet, set the stage for the development and application of artificial intelligence techniques in enterprises in general and financial institutions in particular. In the banking world, its application is even more relevant, thanks to the use of larger and larger data sets for credit risk modelling. The evaluation of credit risk has largely been based on client data modelling; such techniques (linear regression, logistic regression, decision trees, etc.) and data sets (financial, behavioural, sociologic, geographic, sectoral, etc.) are referred to as “traditional” and have been the de facto standards in the banking industry. The incoming challenge for credit risk managers is now to find ways to leverage the new AI toolbox on new (unconventional) data to enhance the models’ predictive power, without neglecting problems due to results’ interpretability while recognizing ethical dilemmas. Contributors are university researchers, risk managers operating in banks and other financial intermediaries and consultants. The topic is a major one for the financial industry, and this is one of the first works offering relevant case studies alongside practical problems and solutions.


Chapter 1. Introduction
Nowadays that new masses of data have become available, and the artificial intelligence (AI) techniques have become more interpretable, the financial service industry is investing on the development of AI models. In particular, alternative types of information have become accessible due to the business relationships of banks with their customers, the progressive digitalisation of the economy, the availability of information of the websites, newspaper articles and social media, the COVID-19 pandemics. Those types of data can be used with different purposes to enhance several aspects of the credit risk modelling: early warning, provisioning, benchmarking, loan granting and risk discrimination.
Rossella Locatelli, Giovanni Pepe, Fabio Salis
Chapter 2. How AI Models Are Built
This chapter describes the various kinds of data that are mostly in use today in AI models, differentiating between “structured”, “semi-structured” and “unstructured” data. Text analysis and Natural Language Processing are illustrated as the main structuring techniques for unstructured data. Some examples of alternative credit data are described, including among others transactional data, data extracted from telephones and other utilities, data extracted from social profiles, data extracted from the world wide web and data gathered through surveys/questionnaires. Also, the chapter describes the opportunity of estimating a model only by means of machine learning techniques, detailing the characteristics of the most used ML algorithms: decision trees, random forests, gradient boosting and neural networks. The application of a special type of neural network is detailed: the autoencoder.
Rossella Locatelli, Giovanni Pepe, Fabio Salis
Chapter 3. AI Tools in Credit Risk
This chapter describes four types of application of AI into Credit Risk modelling. The use of alternative transactional data together with the application of machine learning techniques in the context of the Probability of Default (PD) parameter estimation leads to enhancements of the PD models, able to capture phenomena that were not properly explained by the traditional models. Some examples are described in this paragraph: risk discrimination for borrowers with seasonal business, identification of counterparty risk during the COVID-19 crisis, early warnings and advanced analytics in loan approval Several combinations of traditional modelling techniques and AI techniques can be used to enhance the outcome of the credit risk models. In particular, the business case “two-step approach” is described, detailing the intervention of the AI techniques in a second phase of the model estimation, when the traditional techniques already produced a result. The third part of the chapter describes the application of an AI model to asset management. The model is aimed at supporting an asset manager’s investment decisions. The last section of the chapter describes how to implement machine learning techniques with benchmarking purposes in the context of the validation of credit risk models used for the estimation of the regulatory capital.
Rossella Locatelli, Giovanni Pepe, Fabio Salis
Chapter 4. The Validation of AI Techniques
This chapter describes the implementation of validation techniques aimed at monitoring and mitigate risks related to the development of AI models. The key trustworthy indicators are identified and detailed in coherence with the main trustworthy principles, namely accuracy, robustness, fairness, efficiency and explainability. Also, a focus on the interpretability of the AI models’ outcomes, summarising the main regulatory requirements, and describing the methodological approaches aimed at assessing the stability of the models is detailed. In order to evaluate and interpret the results of the AI models, the contribution of each risk divers is assessed by means of specific methodologies.
Rossella Locatelli, Giovanni Pepe, Fabio Salis
Chapter 5. Possible Evolutions in AI Models
This chapter describes the possible evolution of AI models in the credit risk of tomorrow, evaluating the Regulatory position with respect to the implementation of such techniques with reference to the credit risk assessment, the position of the players in the market, the evaluative economic and macroeconomic environment. Also, an analysis of the outcomes of the AI models is reported, detailing the main aspects concerning ethics, transparency discrimination and inclusion.
Rossella Locatelli, Giovanni Pepe, Fabio Salis
Artificial Intelligence and Credit Risk
verfasst von
Rossella Locatelli
Giovanni Pepe
Fabio Salis
Electronic ISBN
Print ISBN