Key Details
The EAC-PM working paper, Reforms, Efficiency, and Productivity of Indian Banking Sector in the Last Decade: A DEA Approach, examines 12 public, 21 private and 14 foreign banks during FY15–FY26. Together, the selected banks account for over 95% of banking-system assets.
Overall technical efficiency: Increased from 77.99% in FY20 to 88.34% in FY26
Public sector banks: Rose from 72.46% to 93.12%
Private banks: Increased from 78.03% to 86.02%
Foreign banks: Recorded 85.88% in FY26
Full-period averages: 88.5% for PSBs, 85.6% for private banks and 89% for foreign banks
Scale efficiency: Only seven of 47 banks were fully scale-efficient in FY26, compared with 21 in FY15
PSBs Recorded the Strongest Post-FY20 Recovery
The study places the efficiency recovery within a decade of asset-quality recognition, recapitalisation, insolvency reform, bank consolidation, stronger supervision and digitalisation.
Public sector banks (PSBs) recorded the largest improvement after FY20. Their average technical-efficiency score reached 93.12% in FY26, compared with 86.02% for private banks and 85.88% for foreign banks. PSBs also outperformed private banks during FY24–FY26.
The paper associates this recovery with capital infusion, technology upgrades and balance-sheet repair. However, the system-wide FY26 score remained below the 91.29% recorded in FY15, indicating recovery from the FY20 trough rather than continuous improvement across the decade.
Technical Efficiency Measures Resource Use
The paper uses Data Envelopment Analysis (DEA) to assess how efficiently banks convert resources into measurable outputs.
It compares fixed assets, borrowings, operating expenses and employee costs against deposits, advances, investments and net profit or loss.
A score of 100% represents the efficiency frontier within the sample. An 80% score indicates that a bank could theoretically produce the same measured outputs with 20% fewer inputs relative to the benchmark banks.
What Is Technical Efficiency? It measures a bank’s performance relative to other banks using the inputs and outputs selected by the study. It does not directly measure customer service, risk management or wider social value, and results can change with the variables used.
Balance-Sheet Repair Accompanied the Recovery
Conventional indicators show a substantial improvement in banking health:
Gross NPAs fell from 11.5% in March 2018 to 1.68% in June 2026.
Capital adequacy increased from 13% in March 2013 to 17.78%.
Return on assets improved from –0.2% in FY18 to 1.32%.
Return on equity rose from –1.9% to 13.23%.
These measures capture asset quality, capital strength and profitability. DEA addresses a separate question: how efficiently banks used their resources to generate the selected outputs.
Efficiency and Productivity Did Not Always Move Together
The study also uses the Malmquist Productivity Index to distinguish changes in banks’ relative efficiency from shifts in the sector’s technological frontier.
The technology-related measure improved through FY21, fell sharply in FY23 and recovered to 1.0087 in FY26, marginally above the level indicating a positive frontier shift.
A bank can therefore become more efficient relative to its peers without a corresponding sector-wide productivity gain, while technological improvement need not immediately raise its relative efficiency.
The authors identify AI, machine learning and multilingual digital banking as possible sources of future productivity gains, rather than outcomes established by the study.
Consolidation Produced Both Scale and Disruption
Mergers with weaker banks affected the measured performance of some acquiring institutions, while historical data for merged entities were combined to maintain comparability.
Despite higher average efficiency, only seven banks were fully scale-efficient in FY26. Larger size therefore does not automatically place a bank at its most productive operating scale.
The authors favour further consolidation to create several large banks of broadly comparable size. This recommendation goes beyond the efficiency results alone: competition, financial stability, service access and concentration risks would also need to be considered.
Policy Relevance
PSB reform gains appear substantial, but causality remains uncertain. The model measures outcomes before and after a package of reforms; it does not isolate how much improvement came from recapitalisation, mergers, digitalisation, lower bad loans or broader economic conditions.
Efficiency should not become the only performance test. Consumer complaints, credit access, cyber resilience, financial inclusion and lending quality need to accompany cost-and-output measures.
Further mergers require a wider assessment. A higher DEA score alone cannot determine whether consolidation will improve competition, innovation or systemic resilience.
AI investment needs measurable objectives. Banks should specify whether AI is expected to reduce processing costs, improve fraud detection, widen multilingual access or strengthen credit assessment—and monitor the associated bias, privacy and security risks.
Follow the Full Paper Here: Reforms, Efficiency, and Productivity of Indian Banking Sector in the Last Decade: A DEA Approach