Bank Business Models, Size, and Profitability

F. Bolivar,M. A. Duran,A. Lozano-Vivas

DOI: https://doi.org/10.1016/j.frl.2022.103605

2024-01-23

Abstract:To examine the relation between profitability and business models (BMs) across bank sizes, the paper proposes a research strategy based on machine learning techniques. This strategy allows for analyzing whether size and profit performance underlie BM heterogeneity, with BM identification being based on how the components of the bank portfolio contribute to profitability. The empirical exercise focuses on the European Union banking system. Our results suggest that banks with analogous levels of performance and different sizes share strategic features. Additionally, high capital ratios seem compatible with high profitability if banks, relative to their size peers, adopt a standard retail BM.

General Economics

What problem does this paper attempt to address?

This paper explores the relationship between bank size and its business models (BMs) with profitability. The study adopts machine learning techniques, particularly the random forest algorithm, to analyze whether bank size influences its strategic decision-making and identify the business models of banks with different sizes. The paper focuses mainly on the banking systems of the European Union (EU).

The research finds that while the business models of banks vary due to size, banks with similar profitability exhibit similarities in their strategic decisions. Additionally, a high capital ratio is compatible with high profitability, provided that the bank adopts a standard retail business model compared to other banks of the same size. The paper suggests that an increase in regulatory capital requirements does not necessarily conflict with good profit performance, depending on the bank's business model.

The research methodology involves categorizing sample banks into large, medium, and small tiers based on size, and then using the random forest algorithm and tree interpreter to analyze the contributions of various asset and liability portfolios to profitability and identify the business models of banks with different sizes. By using clustering algorithms, the best and worst business models for each size were determined, and significant similarities in strategic decision-making were found among banks with different sizes but similar levels of profitability.

The results of the paper have implications for policymakers, indicating that considering the bank's business model may be more important than size when designing policies to enhance profitability. For example, encouraging banks to move towards a retail business model and reducing derivative exposures can improve the profitability of banks of all sizes. Simultaneously, a high capital ratio can coexist with high profitability, but this depends on the bank's business model.

In conclusion, this paper attempts to address the question of how bank size affects the relationship between its business models and profitability, and how this relationship can guide bank regulation and policy-making.