An Evaluation of Development Bank Efficiency and Its Determinants: A Two-Stage DEA Approach

Purpose: This study evaluates the efficiency of development banks worldwide and identifies its determinants, focusing on whether multilateral development banks differ in efficiency from multilateral development banks.
Research Methodology: Using S&P Global financial data on 98 development banks across five regions from 2010 to 2024, the study applies a two stage approach. Output oriented Data Envelopment Analysis (DEA) under constant returns to scale estimates technical efficiency, and bootstrapped tobit regression examines institutional type, net interest margin, bank size, loan to deposit ratio, nonperforming loans and return on assets as determinants.
Results: The average efficiency score of 0.569 indicates substantial input slack, as banks could reduce inputs by 43.1% without lowering outputs. Multilateral development banks are significantly less efficient than national counterparts, reflecting governance complexity across jurisdictions. Bank size and the loan to deposit ratio have strong positive effects, whereas net interest margin and return on assets are insignificant. Unexpectedly, nonperforming loans are positively associated with efficiency, suggesting that socially oriented lending to riskier borrowers can enhance developmental impact.
Conclusion: Structural balance sheet characteristics, rather than conventional profitability metrics, primarily drive the efficiency of development banks.
Limitation: The analysis relies on a single DEA and tobit framework and on financial indicators of individual banks, without accounting for statistical noise, macroeconomic dynamics or climate and green investments.
Contribution: The study offers evidence from multiple regions comparing multilateral and national development banks within a unified framework, with policy implications for governance reform, liquidity management and scale strategies.