Please rotate your device to landscape mode to view the charts.

Bank financial sustainability evaluation: data envelopment analysis with random forest and Shapley additive explanations

Journal: European Journal of Operational Research

Date: 2025

Author: Yu Shi, Vincent Charles, Joe Zhu

Abstract:
Ensuring financial sustainability is imperative for a financial institution's overall stability. To mitigate the risk of bank failure amid financial crises, effective management of financial sustainability performance becomes paramount. This study introduces a comprehensive framework for the accurate and efficient quantification, indexing, and evaluation of financial sustainability within the American banking industry. Our approach begins by conceptualizing financial sustainability as a multi-stage, multifactor structure. We construct a composite index through a three-stage network data envelopment analysis (DEA) and subsequently develop a random forest classification model to predict financial sustainability outcomes. The classification model attains an average testing recall rate of 84.34 %. Additionally, we employ SHapley Additive exPlanations (SHAP) to scrutinize the impacts of contextual variables on financial sustainability performance across various substages and the overall banking process, as well as to improve the interpretability and transparency of the classification results. SHAP results reveal the significance and effects of contextual variables, and noteworthy differences in contextual impacts emerge among different banking substages. Specifically, loans and leases, interest income, total liabilities, total assets, and market capitalization positively contribute to the deposit stage; revenue to assets positively influences the loan stage; and revenue per share positively affects the profitability stage. This study serves the managerial objective of assisting banks in capturing financial sustainability and identifying potential sources of unsustainability. By unveiling the “black box” of financial sustainability and deciphering its internal dynamics and interactions, banks can enhance their ability to monitor and control financial sustainability performance more effectively.

Link: Google Scholar

Background and Context

Financial Sustainability Challenge

Banks must maintain financial sustainability to prevent failure during financial crises and ensure overall stability.

Multi-Stage Banking Process

Banking operations involve multiple interconnected stages: deposit generation, loan generation, and translation to profitability.

Innovative Analysis Framework

The research combines network Data Envelopment Analysis (DEA), Random Forest classification, and SHAP explanations for comprehensive evaluation.

Three-Stage Banking Process Reveals Complex Financial Sustainability Structure

Deposit Stage Gathering deposits Loan Stage Making loans Profitability Stage Generating profit Inputs: Salary, CapEx, Equity Inputs: Deposits, Assets Inputs: Loans, Interest Income Outputs: Deposits, Cash flow, ROE Outputs: Loans, Interest Income, ROA Outputs: Revenue, EPS
  • The framework models banking as a three-stage process flowing from deposit generation to loan creation to profit generation.
  • Each banking stage has its own inputs, outputs, and performance metrics that contribute to overall financial sustainability.
  • The network structure "unveils the black box" of banking operations to better locate sources of unsustainability.

DEA Scores Reveal Banks Perform Better in Deposit Stage Than Other Stages

  • Banks show strongest financial sustainability performance in the deposit stage with more scores above 0.5.
  • Loan stage and profitability stage show significantly lower performance with majority of scores below 0.2.
  • Only 7 banks (less than 1%) achieved best-practice financial sustainability across all three stages.

Different Variables Drive Financial Sustainability Across Different Banking Stages

Key Contextual Variables by Banking Stage Deposit Stage Loans & Leases + Total Liabilities + Total Assets + Market Capitalization + Loan Stage Total Liabilities +/- Cash from Operations +/- Revenue to Assets + Cashflow per Share + Profitability Stage Revenue per Share + Net Interest Income +/- Total Assets per Share + Net Assets per Share + Symbol meanings: + (positive impact), +/- (mixed impact)
  • Loans and leases, total liabilities, total assets, and market capitalization have strong positive impacts on deposit stage sustainability.
  • Revenue to assets ratio and cashflow per share notably contribute to better financial sustainability in the loan stage.
  • Revenue per share emerges as the strongest driver of financial sustainability in the profitability stage.

Random Forest Model Achieves High Accuracy in Predicting Financial Sustainability

  • The loan stage has the highest prediction accuracy at 91.40%, showing very strong contextual variable relationships.
  • The profitability stage follows with 86.79% accuracy, indicating good predictability from contextual variables.
  • Overall three-stage financial sustainability has the lowest predictability at 69.43%, reflecting its complex nature.

SHAP Analysis Reveals Financial Sustainability Impact Mechanisms

  • Net loans and leases have the strongest positive impact on financial sustainability in the deposit stage.
  • Total liabilities show significant influence across all banking stages, particularly in loan operations.
  • Revenue-related metrics (revenue per share, revenue to assets) are critical to profitability stage sustainability.

Contribution and Implications

  • The framework enables banks to identify sources of unsustainability by decomposing the complex banking process into stages.
  • High prediction accuracy demonstrates that contextual variables can effectively predict financial sustainability performance outcomes.
  • Bank executives can use these insights to monitor critical variables affecting sustainability in different operational stages.
  • The model helps investors identify financially sustainable banks during times of economic uncertainty or financial crisis.

Data Sources

  • Visualization 1 is based on Figure 1 in the paper showing the three-stage DEA model structure.
  • Visualization 2 uses data from Figures 2-5 showing the distribution of DEA scores across banking stages.
  • Visualization 3 draws from SHAP analysis results in Figures 6-8 showing key contextual variables by stage.
  • Visualization 4 is created using data from Table 6 showing testing recall rates of the random forest model.
  • Visualization 5 is derived from SHAP feature importance results in Figures 6a, 7a, and 8a.