Dynamic clustering for business model identification and financial stability


Financial stability is a major concern of financial regulation. This stability may be jeopardized if financial institutions become more and more alike in terms of their business models and strategy. Recent non-standard monetary policies such as ultra-low interest rates for long periods, bond buying programs and the provision of abundant liquidity by central banks have arguably led to pressure on the financial system and convergence of financial institutions’ business models. Little, however, is known about these dynamics of business models in relation to monetary conditions.
This PhD project aims to develop new methodology to investigate the dynamics of financial business models and their convergence/divergence in times of stress and non-standard monetary policies. The project addresses flaws in previous methodology by allowing business model switches in a dynamic, changing environment while retaining model tractability by the use of recent score-driven time series models. Methodologically, the proposal pushes the boundary of time series econometrics into dynamic clustering and machine learning, while conversely we introduce time-varying parameter methodology in the machine learning arena.
The results will be applied to the European banking sector and will be developed in a network that involves prime regulatory players such as the ECB and BIS as well as national central banks. Dissemination of the results will be ensured by providing easy code of these models in currently often-used coding languages (Python, R, Matlab).





Prof. dr. A. Lucas

Verbonden aan

Vrije Universiteit Amsterdam, School of Business and Economics, Department of Econometrics and Operations Research


01/09/2019 tot 01/09/2023