Banked but Not Included: The Account Gap Is Not a Poverty Gap

Executive Summary. Fewer than half of Bangladeshi adults, 43.28%, held an account at a financial institution or with a mobile-money provider in 2024, fifth of six Asian economies and 27.27 percentage points below the peer median of 70.55%. The obvious explanation, that a poorer country banks fewer people because fewer people have money worth banking, does not survive the distributional cuts. Bangladesh's richest 60% are banked at 48.19%, which is beneath the peer median for the poorest 40% (48.42%) and stands at about 56% of the 85.49% the median comparator reaches in its own top three-fifths. The poorest 40%, at 35.55%, sit at about 73% of their peer median. The gap to the region is proportionally wider at the top of the Bangladeshi distribution than at the bottom. It is wider still among the young: 38.73% for adults aged 15-24, 4.56 points under Bangladesh's own adult rate, where in the median comparator the young sit above it (72.16% against 70.55%). None of this is a handset shortage. Bangladesh ran 108.1 mobile subscriptions per 100 people in 2024, fourth of six.

Fifth of six, and 27 points short

The 2024 readings, in order: Thailand 91.82%, India 89.02%, Vietnam 70.55%, Indonesia 56.33%, Bangladesh 43.28%, Pakistan 27.3%. Bangladesh is fifth of the six, above only Pakistan, sitting 27.27 percentage points under the peer median of 70.55% and at about 61% of it.

Note what the indicator counts and what it does not. It counts an adult who reports holding an account at a bank or another financial institution, or with a mobile-money provider. Holding. Not using, not saving in, not borrowing against. The World Bank series behind this essay carry no transaction, balance or dormancy data for any of the six countries, so nothing here separates a live account from one opened once and abandoned. That limitation applies to Bangladesh's figure and to every comparator's equally.

Source: World Bank World Development Indicators, account ownership at a financial institution or with a mobile-money provider (FX.OWN.TOTL.ZS), share of adults ages 15 and over, 2024.

The gap is widest at the top of the distribution

If household income were the binding constraint, exclusion would concentrate at the bottom and the top would look regional. In Bangladesh the top does not look regional.

Among the richest 60%, ownership is 48.19% against a peer median of 85.49%: a gap of 37.3 percentage points, or about 56% of the median. Among the poorest 40% it is 35.55% against a peer median of 48.42%: a gap of 12.87 points, or about 73% of the median. Bangladesh is fifth of six on both cuts, ahead of Pakistan on each.

Two comparisons make the shape concrete. Vietnam's poorest 40% are banked at 48.11%, which is where Bangladesh's richest 60% sit (48.19%). India's poorest 40%, at 86.32%, are far above Bangladesh's richest 60%. The Bangladeshi profile is flat and low, 48.19% at the top against 35.55% at the bottom, where Vietnam runs 85.49% against 48.11% and Pakistan 36.41% against 13.61%.

A flat, low profile is what an obstacle that applies to everybody produces, rather than one that scales with poverty. The candidates are familiar: documentation and identity checks at the moment of opening, distance to an authorised agent or branch, the cost and design of the product itself, and whether wages, remittances and public transfers arrive in an account by default or in cash. A six-country panel cannot test any of them. What it can do is discount the simplest story, because the Bangladeshis with the most money are the ones sitting furthest from their regional counterparts.

One qualification, in the other direction. The poorest-40 peer median of 48.42% is itself a low bar, so Bangladesh's narrower relative gap at the bottom partly reflects that most comparators also under-bank their poor. India (86.32%) and Thailand (91.19%) are the exceptions, and Vietnam, at 48.11%, is not.

The young sit behind their own country, and women further out again

Among adults aged 15-24, ownership is 38.73%, which is 4.56 percentage points below the all-adult rate of 43.28%. Bangladesh is fifth of six here too, above only Pakistan's 24.34%, and at about 54% of the peer median of 72.16%, a proportionally wider gap than for adults as a whole.

The direction inside the country matters as much as the level. In two of the five comparators the young are banked at a higher rate than adults overall: Indonesia 61.99% against 56.33%, Vietnam 72.16% against 70.55%. In three of them, as in Bangladesh, they are not: India 83.33% against 89.02%, Thailand 88.13% against 91.82%, Pakistan 24.34% against 27.3%. A younger cohort below the adult average is therefore not peculiar to Bangladesh, and some of the 15-24 shortfall anywhere is students and dependants with little to put in an account. What is distinctive is the distance: at 38.73%, the Bangladeshi cohort now entering the labour market is proportionally further from its regional counterparts than the adult population is from its own.

Women are further out again. Female ownership is 33.35%, fifth of six, above only Pakistan's 11.91%, against a peer median of 69.88%: a gap of 36.54 points, and about 48% of the median. Of the five ownership cuts in this data, women sit furthest below their peer median in proportional terms. Thailand reaches 92.7% among women, India 89.25%, Vietnam 69.88%, Indonesia 58.03%.

Source: World Bank WDI, account ownership among young adults ages 15-24 (FX.OWN.TOTL.YG.ZS), 2024.

The handsets arrived, the accounts did not

Bangladesh ran 108.1 mobile subscriptions per 100 people in 2024, more than one per person, fourth of six and at about 88% of the peer median of 122.5. In 2000 the figure was 0.2074 per 100 people. The subscription base is 521.2 times its level at the turn of the century.

Set that against ownership of 43.28% and the distance between a SIM and an account is what needs explaining.

Within this group, subscription density does not order account ownership. India runs 79.35 subscriptions per 100 people and Pakistan 76.85, both under Bangladesh's 108.1, and India's account ownership is 89.02%. Bangladesh has more connections per head than India, and an account rate of 43.28% against India's 89.02%.

Two caveats on the subscription series. It counts active SIMs, not people, and multi-SIM holding inflates every country's reading; Thailand's 160.6 per 100 people is not a claim about how many handsets a Thai adult carries. The series is an upper bound on reach, not a measure of individual access. And a SIM and an account are not issued under the same rules: a connection is sold quickly and cheaply, while an account requires identity documents, verification, an agent or branch authorised to onboard, and a product the holder judges worth keeping. The friction sits in that second list, and this data cannot say which item in it binds hardest.

Connectivity does not order accounts either. 53.42% of Bangladesh's population used the internet in 2024, lowest of the six, 19.36 percentage points under the peer median of 72.78% and at about 73% of it. Thailand reached 90.87%, Vietnam 84.15%, Indonesia 72.78%, India 64.94%, Pakistan 57.25%. Yet Pakistan sits above Bangladesh on internet use and far below it on account ownership, at 27.3%.

That reading constrains app-based finance rather than mobile money as such, since basic services run over menu codes on feature phones and need no data connection. The internet series is also a share of the whole population while the ownership series is a share of adults, so the two cannot be stacked. What the pair establishes is narrow and still useful: Bangladesh is mid-table on subscriptions and at the bottom of the group on internet use, and neither reading predicts where it lands on accounts.

Source: World Bank WDI, mobile cellular subscriptions per 100 people (IT.CEL.SETS.P2), Bangladesh against comparators, 2000-2024.

What an ownership count cannot tell you

Four limits, stated before anyone else states them.

This is a cross-section of six countries in a single year. The associations in it identify no cause, and six countries is a comparison rather than a sample. The levels are survey-based and self-reported, so two decimal places describe the source, not the precision of the world.

There is no time series for ownership in this data, only the 2024 cross-section, so nothing here says whether Bangladesh's gap is closing or widening.

Whether any comparator's level came from a national enrolment drive, from the digitisation of wage and transfer payments, or simply from income growth is not something these seven indicators can distinguish. A high peer number is a destination, not a route.

And the second half of this essay's title is a hypothesis the data cannot confirm. Ownership counts holding, so it cannot show how many of the 43.28% transact, save or borrow. What it does fix is a ceiling: inclusion cannot exceed the count, and the count is 43.28%.

What would change this conclusion

A specific, checkable test. The claim here is that the constraint is not household income, which implies the top of the distribution can move without waiting for incomes to converge. If the next comparable vintage puts Bangladesh's richest 60% at or above Indonesia's current 61.59% while the poorest 40% stay near 35.55%, the binding constraint was registration and reach, relieved first for the people easiest to reach. If the richest-60 rate is instead still near 48.19% after another stretch of output growth at rates like 2024's 4.22%, the same conclusion arrives by a different route: incomes rose and ownership did not follow. The reading that falsifies the essay is the third one, both cuts rising together and roughly in proportion to incomes, which would mean the account gap was a poverty gap all along.

Three moves, with owners and signals.

1. Make the account the default destination for money that already moves. Wages in registered enterprises, public transfers and utility collections landing in an account held in the recipient's own name convert flows that already exist into ownership, without requiring anybody to go looking for a bank. Owner: Bangladesh Bank, with the Financial Institutions Division on scheme design. Success signal: account ownership above Indonesia's current 56.33% in the next comparable vintage, which would move Bangladesh off fifth position among the six.
2. Treat the female gap as the binding one. At 33.35% against a peer median of 69.88%, women are the cut furthest below the region in proportional terms, and no path to a national rate near the peer median of 70.55% runs around them. Owner: Bangladesh Bank on agent-network and onboarding rules, with BBS reporting outcomes by sex. Success signal: female ownership above the current all-adult national rate of 43.28%.
3. Publish account activity, not just account holding. An ownership count cannot separate a working account from a dormant one, and the whole banked-but-not-included question turns on that separation. Owner: Bangladesh Bank, in the regular payment-systems statistics. Success signal: a published series on the share of accounts with no transaction in a full quarter, split by sex and account type, so the count and its use can be read together.

The counterargument

The strongest objection is a measurement one and it is not weak. Bangladesh moves a large volume of money through mobile financial services over the counter: a customer hands cash to an agent, the agent sends it from a registered account, the recipient collects cash at the other end. Nobody in that chain needs an account of their own. On this reading, 43.28% counts registrations rather than access, and the country scores badly because it built a service model that does not require universal registration.

Three answers.

First, the objection is right that this data cannot rule it out. There are no transaction figures in these indicators and the essay does not pretend otherwise.

Second, over-the-counter access and inclusion are different goods. An account in your own name holds a balance, accumulates a savings and repayment record, and gives its holder a claim they can exercise. Transacting through somebody else's account delivers the transfer and none of the rest, and the rest is the part that compounds over a working life.

Third, and this is where the objection stops reaching: the distributional pattern. If the count merely mismeasured a service-based model, the mismeasurement should not be concentrated among the richest 60%, who are the group best placed to satisfy documentation requirements and most likely to register in their own name. Their 48.19% against a peer median of 85.49% is the fact that has to be explained, and an agent-based service model does not explain it.

Six countries in one year is a comparison, not an experiment, and nothing here identifies a cause. What the comparison does establish is a shape: a shortfall that fails to taper as you move up the income distribution, which is not the shape a poverty explanation predicts.

---

Data sources: World Bank World Development Indicators, retrieved from the BDPolicyLab data lake, 2026-08-10. Comparators are India, Pakistan, Indonesia, Thailand and Vietnam; "peer median" is the median of the five comparators, excluding Bangladesh. All six countries report every indicator used here, so each ranking rests on six observations. Account ownership, mobile subscriptions and internet use are 2024 readings; the mobile series runs from 2000. Ownership covers adults ages 15 and over except the young-adult cut, which covers ages 15-24; the richest-60 and poorest-40 cuts are ownership rates among adults in those household income groups.

Sources

1. World Bank WDI, Account ownership at a financial institution or mobile money (% of pop 15+), FX.OWN.TOTL.ZS: https://data.worldbank.org/indicator/FX.OWN.TOTL.ZS
2. World Bank WDI, Account ownership, young adults 15-24 (% of pop), FX.OWN.TOTL.YG.ZS: https://data.worldbank.org/indicator/FX.OWN.TOTL.YG.ZS
3. World Bank WDI, Account ownership, female (% of pop 15+), FX.OWN.TOTL.FE.ZS: https://data.worldbank.org/indicator/FX.OWN.TOTL.FE.ZS
4. World Bank WDI, Account ownership, richest 60% (% of pop 15+), FX.OWN.TOTL.60.ZS: https://data.worldbank.org/indicator/FX.OWN.TOTL.60.ZS
5. World Bank WDI, Account ownership, poorest 40% (% of pop 15+), FX.OWN.TOTL.40.ZS: https://data.worldbank.org/indicator/FX.OWN.TOTL.40.ZS
6. World Bank WDI, Mobile cellular subscriptions (per 100 people), IT.CEL.SETS.P2: https://data.worldbank.org/indicator/IT.CEL.SETS.P2
7. World Bank WDI, Individuals using the Internet (% of population), IT.NET.USER.ZS: https://data.worldbank.org/indicator/IT.NET.USER.ZS