The Architecture of Power in Financial Institutions

Centralization vs Decentralization of Decisions

A situation familiar to many bankers unfolds more often than we realize.

A long-standing MSME customer seeks a modest enhancement to his working capital limit. The business is stable. The relationship manager knows the borrower well. The regional team supports the proposal.

Yet the file begins its journey.

Branch → Regional Office → Head Office → Credit Department → Committee.

Clarifications are raised. Additional data is sought.Internal discussions continue.

By the time approval finally arrives, the borrower has already arranged funds elsewhere.

Nothing in the system has failed.

Policies were followed. Controls worked exactly as designed.

And yet something important becomes clear.

The real determinant of decision speed in financial institutions is not technology or policy manuals.

It is where decision power sits.

Across banks and NBFCs, one can broadly observe five distinct architectures of decision-making power.

Two Forces That Shape Decision Power

The design of decision-making structures in financial institutions is shaped by two competing forces:

Speed of decision-making and degree of risk control.

Institutions that prioritize control often centralize authority. Institutions that prioritize responsiveness distribute authority or automate decisions.

This tension creates different institutional models.

Decision Architecture Matrix

Model 1: The Centralized Control Model

(High control, slower decisions)

In this structure, most meaningful decisions are concentrated at Head Office.

Branches originate business and conduct initial analysis, but final authority rests centrally with credit teams or committees.

How it works

A proposal may move through multiple levels:

Branch → Regional Office → Head Office Credit → Committee → Final Approval

Each level introduces additional review and oversight.

Example

Even a modest working capital enhancement for an MSME borrower may require:

• Head Office credit review

• policy validation

• committee approval.

  
  

The objective is clear: uniform policy enforcement and strong centralized control.

Trade-off

While the structure strengthens governance, it can also create decision latency, as authority sits far from the customer.

📦 Illustrative Institutions

• Bandhan Bank

• Yes Bank (during post-restructuring tightening)

• South Indian Bank ( Of late, moving towards Model 2)

• Several large Indian Public Sector Banks

• Bajaj Finance

• Mahindra & Mahindra Financial Services

• L&T Finance

These institutions emphasize centralized credit governance and standardized policy interpretation.

Model 2: The Structured Delegation Model

(High control, faster decisions)

Many large private sector institutions operate under a more balanced structure.

Policies remain centralized, but decision authority is distributed through defined sanction limits.

How it works

Approval authority is tiered across the organization:

• Branch authority – smaller exposures

• Regional committee – mid-sized exposures

• Zonal committee – larger exposures

• Head Office – strategic or large-ticket exposures.

  
  

Example

A ₹3 crore working capital facility for a manufacturing firm may be approved by a regional credit committee, without requiring Head Office approval.

Trade-off

This model preserves policy discipline while improving decision speed and regional accountability.

📦 Illustrative Institutions

Institutions widely seen operating under structured delegation frameworks include:

• ICICI Bank

• HDFC Bank

• Kotak Mahindra Bank

• Axis Bank

  
  

These institutions combine regional decision authority with centralized risk monitoring systems.

Model 3: The Relationship Banker Model

(Judgement-driven decision architecture)

Some institutions historically evolved around a different philosophy: trust in experienced bankers.

Instead of relying heavily on committees and layered approvals, they relied on deep borrower familiarity and local knowledge.

Example

A senior regional banker who has followed a textile cluster for decades may understand:

• promoter credibility

• seasonal working capital cycles

• local industry dynamics.

Such contextual insight allows decisions to be taken quickly and intelligently.

Trade-off

The model produces strong customer relationships but depends heavily on institutional culture and quality of people.

📦 Illustrative Institutions

Institutions historically associated with strong relationship-driven cultures include:

• Sundaram Finance

• City Union Bank

• Karur Vysya Bank

• Federal Bank (in SME segments)

  
  

These institutions traditionally relied on experienced local bankers with deep borrower familiarity.

Model 4: The Algorithmic / Hybrid Decision Model

(Technology-driven lending)

The rise of fintech has introduced algorithm-driven lending.

In this structure, credit decisions are based largely on rule engines and data analytics.

How it works

Applications are evaluated through automated analysis of:

• credit bureau data

• bank statements

• transaction analytics

• digital payment histories.

If parameters fall within policy thresholds, approvals can occur within minutes.

Trade-off

Algorithmic lending enables speed and scale, but introduces model risk when algorithms fail to capture contextual nuances.

📦 Illustrative Institutions

• ICICI Bank

• HDFC Bank

• Axis Bank

• Kotak Mahindra Bank

• Lendingkart

• Capital Float

• NeoGrowth

Model 5: The Data-Network Decision Model

(The emerging architecture)

A new model is now taking shape in India, driven by national digital infrastructure.

Instead of relying only on internal data or algorithms, lending decisions increasingly draw from large national data networks.

These include:

• GST data

• Account Aggregator financial information

• digital payment histories

• income tax data.

• 
  

Example

A small business applying for credit may be evaluated through:

• GST turnover analytics

• bank transaction cash flows

• bureau data

• account aggregator financial statements.

This creates a much richer and more verifiable credit picture.

Why this matters

India’s digital public infrastructure is quietly reshaping credit assessment.

Decisions are no longer based solely on:

• centralized committees

• local judgement

• algorithmic models.

They are increasingly based on shared national financial data ecosystems.

📦 Illustrative Institutions

Banks and platforms building around this emerging model include:

Banks

• ICICI Bank

• HDFC Bank

• Axis Bank

Data and fintech ecosystems

• Perfios

• Razorpay

• Open Financial Technologies

The Real Lesson

Successful financial institutions rarely operate under a single model.

Instead, they combine elements of several:

• centralized policy frameworks

• distributed decision authority

• technology-driven analytics

• data-network ecosystems.

The challenge is not choosing between centralization and decentralization.

The challenge is designing the right architecture of power.

A Final Thought

In finance, risk does not arise only from bad decisions.

Sometimes risk arises from decisions that arrive too late.

The institutions that succeed will not simply control risk.

They will design systems that allow them to control risk while still moving fast enough to serve the real economy.