Competition between FinTech and Traditional Banks: Evidence from the French Banking Industry  (With Frederic Jouneau-Sion (Univ Lyon 2)).

Abstract: The rapid expansion of FinTech has raised a central question in banking: who will  survive the competition between fully remote banks and traditional networks?This paper develops a spatial competition model in which a network of physical agencies competes with a fully remote provider across heterogeneous local markets. The framework highlights how the uniform pricing constraint across markets and the importance of geography jointly determine equilibrium market coverage and profitability. Using three original geolocated datasets covering 15 million individuals, over 16,000 bank agencies, and over 100,000 customers of a French payment FinTech, we estimate the key parameters of the model. The results indicate that, under estimated levels of consumer reluctance, a fully remote banking model would struggle to reach profitability, even at zero markup. These findings provide quantitative evidence on the limits of the so-called disruptive nature of digital banking.

Working paper — October 2026.

Codifying Trust: Formal Classification and Relational Hierarchy in Decentralized Autonomous Organizations (With Magali Chaudey (Univ Saint Etienne)).

Working paper — September 2026.

Abstract: Blockchain technology was designed to enable coordination without relying on trusted third parties. Yet blockchain-based organizations, known as decentralized autonomous organizations (DAOs), rely on formal classification mechanisms that translate participants' past behaviour into publicly observable levels of trust. We examine whether these formal trust classifications correspond to actors’ positions within DAO governance networks. We analyse three major decentralized finance DAOs that are MakerDAO, Aave, and Uniswap using data on 18,735 actors and social network analysis of governance interactions. We find that trust classification and network position are positively associated, but the two dimensions do not fully coincide. This association persists after accounting for participation volume, and the correspondence is more heterogeneous at the upper levels of the classification. At the highest level, promotion depends on staff judgment rather than behavioural criteria. The findings suggest that behavioural transparency and algorithmic classification do not eliminate the differences in actors’ relational positions. Rather, formal recognition and network position can develop alongside each other, creating forms of organizational standing that are not fully captured by formal classification.