Measuring Financial Risks: The Application of Network Theory in Fintech Risk Management

Recent advancements, gradually transforming the traditional economic and financial system, are mainly characterized with the emergence of digital-based systems. Such systems present a paradigm shift from traditional infrastructural systems to technological (digital) systems. Financial technological (Fintech) companies are gradually gaining ground in major developed economies across the world. The emergence of Peer-to-Peer (P2P) platforms is a typical example of a Fintech system. The P2P platform aims at facilitating credit services by connecting individual lenders with individual borrowers without the interference of traditional banks as intermediaries. Despite the various advantages, P2P systems inherit some of the challenges of traditional credit risk management. In addition, they are characterized by the inability to solve for asymmetric information as efficiently as banks and by differences in risk ownership which in turn might motivate them to push volume even in view of reduced credit standards. Finally, P2P systems note a strong interconnectedness among their users which makes distinguishing healthy and risky credit applicants difficult, thus affecting credit issuers. There is, therefore, a need to explore methods that can help improve credit scoring of individual or companies that engage in P2P credit services. We argue that P2P platforms, through the use of non-traditional data sources as well as advance modelling, can offer a new approach on credit risk evaluation in the context of P2P systems. Specifically, we suggest that the use of alternative data that summarize the interconnections that emerge between borrowers could counterbalance the inherent risks of the business model and in turn lead to higher accuracy in risk classes assignment. Namely, P2P systems can benefit from the inclusion of information on the interconnections or similarities that emerge between different participants on the platform, i.e can benefit from the application of network theory in the credit risk evaluation. Consequently, the overall objective of this thesis is to test the predictive utility of traditional credit scoring models as they are employed in the context of P2P systems and investigate whether the inclusion of network parameters i.e. information on how borrowers are connected, can improve the predictive utility of models. In this work, we propose several approaches on how network theory can be employed to improve the statistical-based credit scoring for P2P systems and those are: (i) correlation-based credit scoring (in the case in which time-varying financial information on borrowers is available on the platform); (ii) similarity-based credit scoring (for cross-sectional data), (iii) factor-network-based segmentation. Furthermore, the thesis also includes an application of network theory in improving Fintech risk management, in a context beyond Fintech credit. Specifically, we also provide an application of network theory in understanding the dynamics of Bitcoin blockchain trading volumes and, specifically, how different trading groups, in different geographic areas, interact with each other. The empirical results presented in this thesis suggest that credit risk management of SMEs engaged in P2P credit services can be improved by employing network theory. Specifically, we demonstrate the effectiveness of our approach through empirical applications analyzing the probability of default of several different samples of SMEs involved in P2P lending across Europe. In each case, we compare the results from our network-augmented model with the one obtained with standard credit score methods and throughout we find that the network-based methodologies lead to an improvement in predictive utility. This finding further remains valid also in the context of alternative P2P systems i.e. the Bitcoin network. We find that our network-based model for understanding the dynamics of trading volumes, overperforms a pure autoregressive model.