How ATMs, Bank Branches, and Mobile Money Agents Define Financial Access Infrastructure

Introduction

When a major earthquake struck Japan on July 29, 2026, the country’s financial access infrastructure faced an immediate stress test. Japan has a relatively high 109.6 ATMs per 100,000 adults (ranked 17th globally), but for a population accustomed to near-universal digital payments (95.8% of adults), the disruption raised a fundamental question: which type of financial access point is most resilient when disaster hits?

The answer varies by economy. Financial access infrastructure rests on three main pillars — ATMs, bank branches, and mobile money agents — and each economy has a different mix. This article explains how the three types compare, which economies rely on which model, and what happens when the system is put under stress.

The Three Pillars of Financial Access

1. ATMs: The Cash Infrastructure

ATMs provide the most basic financial access point: cash withdrawal. They are capital-intensive to deploy (each machine costs thousands of dollars) and require electricity, network connectivity, and regular cash replenishment. But they are also the most automated access point, requiring no human teller.

Where ATMs dominate: High-income economies with extensive banking infrastructure. Uruguay leads the world with 314.8 ATMs per 100,000 adults, followed by Macau (259.6) and South Korea (257.1). Most Western European economies have 80-170 ATMs per 100,000 adults.

The disaster resilience question: Modern ATMs often have backup battery power and can operate for hours during a blackout. In Japan’s earthquake, convenience store ATMs (which have generator backup and are distributed across thousands of 24-hour locations) became critical access points. The main vulnerability is network dependence — if cellular or wired networks go down, even ATMs with power cannot process transactions.

2. Bank Branches: The Full-Service Model

Bank branches offer the widest range of services — deposits, withdrawals, loans, account opening, financial advice — but at the highest cost per transaction. Branch density has been declining globally, with 110 of 172 economies losing bank branches over the past decade.

The density gap: In high-branch-density economies like San Marino (122 branches per 100k adults) or Bulgaria (92), residents can walk into a branch within minutes. In low-density economies like Nigeria or Kenya, there may be only one branch for every 50,000-100,000 adults — compared to one mobile money agent for every few hundred.

The disaster resilience question: Branches require staff to be present and operational, making them vulnerable to transport disruption, staff shortages, and building damage during disasters. However, branches can serve as coordination hubs for emergency financial services — many Japanese banks activated disaster-response protocols immediately after the earthquake, extending branch hours and waiving fees.

3. Mobile Money Agents: The Lean, Distributed Model

Mobile money agents — local retailers, kiosks, or small businesses that process cash-in/cash-out transactions — are the fastest-growing access infrastructure. They are far cheaper to deploy: a mobile money agent needs little more than a basic phone and float capital.

Where mobile money agents dominate: Sub-Saharan Africa leads. Kenya has one of the highest agent densities in the world, supporting 87.5% mobile money account penetration despite having only 6.6 ATMs per 100,000 adults. Nigeria and Ghana have similarly agent-heavy models.

The disaster resilience question: Mobile money agents depend on the mobile network. If cell towers go down (as they can during earthquakes), agents cannot process transactions. However, the distributed nature of the agent network is a strength — thousands of small access points are more resilient than a few concentrated bank branches. Mobile money systems have proven valuable during crises in East Africa and Southeast Asia, as covered in our analysis of how mobile money helps during natural disasters.

The Three Models Compared

Access ModelCapital CostTransaction CapacityReach (last mile)Disaster Vulnerability
ATMsHighHigh (unlimited per machine)ModerateNetwork + power dependent
Bank BranchesVery highHigh (staffed)LowStaff + building dependent
Mobile Money AgentsLowLow (per agent)Very highMobile network dependent

What the Data Shows

The three pillars coexist in different proportions depending on a country’s development level, geography, and regulatory environment:

High-income, high-density economies like Japan, Canada, and South Korea have high density across all three pillars. They combine many ATMs with many bank branches and growing digital payment adoption. During a disaster, these economies can fall back on redundancy — if one pillar fails, another may still function.

Emerging economies with mobile-first models like Kenya and Nigeria have very low ATM and branch density but high mobile money agent density. Their financial access depends on the resilience of the mobile network. For a deeper look at the agent model, see our ranking of mobile money agent density.

Low-infrastructure economies like South Sudan and Sierra Leone have low density across all three pillars. With fewer than 3 ATMs per 100,000 adults and minimal mobile money agent coverage, financial access is extremely limited even in normal times — and essentially absent during emergencies.

Which Model Is Most Resilient?

There is no single answer. In the Japan earthquake, ATMs (especially convenience-store machines) provided the most immediately available cash access because they are distributed and have backup power. Bank branches, while most vulnerable to physical damage, serve as coordination hubs. Mobile money, though absent at scale in Japan, has proven its resilience in Sub-Saharan African crises.

The most resilient financial access infrastructure is one with redundancy across all three pillars — a lesson that both high-income and developing economies can learn from the Japan earthquake.

For a full ranking of ATM density across 174 economies, see our ATM density rankings 2026. For the bank branch perspective, see our analysis of where banking still means walking in. And for the mobile money view, see our article on bank branches vs. mobile money agents.

Sources & method

The ATM density data is from the World Bank’s World Development Indicators (indicator code FB.ATM.TOTL.P5). Bank branch data is from the same source (indicator code FB.CBK.BRCH.P5). Mobile money agent data comes from the IMF Financial Access Survey. Financial inclusion indicators (account ownership, digital payments, mobile money accounts) are from the World Bank Global Findex survey. For the full methodology, see our methodology page.