PhD Defence Lennart John Baals | Risk Management in Digital Finance: Assessment and Pricing in an Emerging Fintech Era

Risk Management in Digital Finance: Assessment and Pricing in an Emerging Fintech Era

Lennart John Baals is a PhD student in the department Industrial Engineering & Management Science. (Co)Promotors are dr. J.R.O. Osterrieder and prof.dr.ir. M.R.K. Mes from the faculty of Behavioural Management and Social Sciences (BMS), University of Twente and prof.dr. A. Hirsa from the Department of Industrial Engineering and Operations Research, Columbia University, New York City.

This thesis investigates how investment risk can be measured and mitigated in digital finance, focusing on digital token markets and peer-to-peer (P2P) lending platforms. As financial services increasingly move to software-mediated platforms, traditional risk tools often fall short in addressing the novel market structures these platforms introduce. The research designs interpretable, statistically disciplined methods for screening, pricing, and surveillance of tokenized assets and P2P credit.

The thesis comprises four empirical studies spanning market risk, credit risk, and systemic risk. The first study applies recursive unit-root tests and log-periodic power-law models to NFT and DeFi token markets, providing some of the earliest systematic detection of speculative price bubbles in these markets. The diagnostics successfully date-stamp boom and bust cycles during 2020-2022, with flagged exuberant windows followed by materially larger drawdowns.

The second study constructs a network-based credit scoring framework for P2P loans. By building similarity graphs over borrower characteristics and computing network centrality metrics such as PageRank, the approach delivers consistent gains in default prediction, up to ten percent improvement in AUC, across multiple classifier architectures.

The third study examines whether interest rates on P2P platforms adequately compensate for risk. Using community detection algorithms to segment borrowers into risk-homogeneous groups, the analysis reveals economically meaningful mispricing that diminishes over time, consistent with a platform learning effect.

The fourth study links P2P loan pricing to macroeconomic conditions through a regime-aware framework combining robust principal-component analysis and hidden-Markov models. Restrictive macro regimes are associated with higher required risk premia, particularly for longer maturities and riskier credit grades.

Together, these contributions provide a practical foundation for building faster, more inclusive, and more resilient digital financial markets.