Paid at the Counter: The State That Left Its Own Health System

Executive Summary. Bangladesh runs the most privatised health financing model of any large Asian economy in its comparison group, and it is not the result of a policy choice anyone defends out loud. Domestic general government health expenditure was 0.31% of GDP in 2023, last of six countries and roughly a fifth of the peer median of 1.45%. The state funds 14.5% of the country's health spending; households pay 79.3% directly at the point of care, the highest out-of-pocket share in the group and more than double the peer median of 39.2%. Translated to money, the government spends US$7.74 per person per year on health, while households hand over US$42.40 each. That second figure is the one that should end the debate: Bangladeshi households pay more out of pocket per person than Thai households do (US$32.44), inside a system that spends a sixth as much in total. And yet Bangladeshi men have the longest life expectancy of the six, 73.3 years in 2024. That combination is real, and it is about to stop working, because the gains it rests on came from interventions that never needed a hospital, and the disease burden now arriving does.

The number that does not look real

Start with the figure that most readers will assume is a typo.

In 2023, domestic general government health expenditure in Bangladesh was 0.31% of GDP. The decimal point is in the right place. The comparison group: Thailand 3.53%, Vietnam 2.02%, Indonesia 1.45%, India 1.30%, Pakistan 0.90%. Bangladesh is last, and the gap is not marginal. The peer median is 1.45% of GDP, so the Bangladeshi state commits about one fifth of what a median comparator commits, measured against the size of its own economy.

Total health spending, from all sources combined, was 2.17% of GDP, also the lowest of the six against a peer median of 3.34%. So the shortfall is not that the state spends little while the private sector spends lavishly. The whole system is small, and within that small system the state's share is smaller still: 14.5% of current health spending is government money. In Thailand the equivalent share is 77.9%.

Source: World Bank World Development Indicators, domestic general government health expenditure (SH.XPD.GHED.GD.ZS) and current health expenditure (SH.XPD.CHEX.GD.ZS), 2023.

This is worth stating plainly because the usual framing gets it backwards. Bangladesh is not a country whose government spends modestly on health. It is a country whose government has very nearly withdrawn from health financing, leaving a system that runs on cash carried to the counter.

What that means per person

Shares of GDP are abstract. Per person, the picture sharpens.

Current health expenditure was US$53.46 per person in 2023. Of that, the government's 14.5% share comes to US$7.74 per person per year. Out-of-pocket payments, at 79.3%, come to US$42.40 per person per year.

Now set that beside Thailand, which is the group's success story in health financing. Thailand spent US$326.50 per person, six times the Bangladeshi total. But because the Thai state carries 77.9% of it, Thai households paid only 9.9% out of pocket, which works out to US$32.44 per person.

So the Bangladeshi household pays US$42.40 and the Thai household pays US$32.44. The poorer country's families pay more cash per head for health care than the richer country's families do, out of far smaller incomes, and receive a far smaller system in return. Bangladesh has not built a cheap health system. It has built an expensive one for the people using it, and a cheap one for the treasury.

Source: World Bank WDI, out-of-pocket expenditure as a share of current health expenditure (SH.XPD.OOPC.CH.ZS) and current health expenditure per capita (SH.XPD.CHEX.PC.CD), 2023.

Out-of-pocket financing has a well-understood property: it charges the sick rather than the well, and it charges them at the exact moment their earning capacity is compromised. A financing system with these shares does not spread risk. It concentrates it, on households at their least able to bear it.

The result nobody predicts

Here is where the data refuses to behave.

Bangladeshi men had a life expectancy at birth of 73.3 years in 2024, the highest of the six countries. India 70.7, Vietnam 70.1, Indonesia 69.2, Pakistan 65.5, and Thailand, the country with eleven times the public health commitment, 72.3. Bangladesh sits 3.3 years above the peer median.

The long view is more striking still. Bangladeshi life expectancy at birth, both sexes, rose from 55.8 years in 1990 to 74.9 in 2024, a gain of 19.1 years, or 34.2%, in a single generation.

Source: World Bank WDI, life expectancy at birth, total (SP.DYN.LE00.IN), Bangladesh against comparators, 1990-2024.

Two honest caveats before anyone builds a policy on that comparison.

First, the male ranking is not matched on the female side. Bangladeshi women live 76.7 years, third of the six, behind Thailand's 81.0 and Vietnam's 79.4. Bangladesh leads the group on male life expectancy and trails the leaders on female life expectancy, which means part of the headline is a statement about male mortality elsewhere, not only about Bangladeshi success. Thailand's male figure carries burdens (road injury and tobacco chief among them) that have nothing to do with the financing of its hospitals.

Second, life expectancy is a stock, not a flow. The 2024 figure reflects decades of accumulated survival, so it reports on the health system of the 1990s and 2000s at least as much as on the one operating now. And it is a modelled series: the Bangladesh Bureau of Statistics' own SVRS produces a different level on a different base, and the two should not be spliced.

Why it worked, and why that ends

The reconciliation is not mysterious once you ask what actually produced the gains.

Bangladesh's mortality decline was driven by interventions that are cheap, one-off or near enough, deliverable outside a clinic, and forgiving of weak follow-up: immunisation, oral rehydration, vitamin A, tetanus toxoid, a collapse in fertility that removed high-parity births, and slow improvements in water and sanitation. None of these require hospital beds, specialists, diagnostic imaging, or continuity of care. A state can buy a great deal of survival this way on 0.31% of GDP, especially with a dense NGO delivery network and a population living close together, which lowers the cost of reaching people.

The delivery capacity confirms the pattern. Bangladesh has 0.68 physicians per 1,000 people (2021), below the peer median of 0.90, and 0.94 hospital beds per 1,000 (2021), the lowest of the four countries reporting, against a median of 1.59. This is not the infrastructure of a country that bought its life expectancy with clinical medicine. It is the signature of a country that bought it with public health.

That model is now running out of road, for a reason internal to its own success. Populations that stop dying of infectious disease and childbirth start dying of cardiovascular disease, diabetes, chronic respiratory disease, and cancer. Those conditions invert every property that made the Bangladeshi model cheap. They require lifelong medication rather than a single dose, continuous monitoring rather than a campaign, specialist diagnosis rather than a community worker, and they punish interruption. They are exactly the conditions that out-of-pocket financing handles worst, because the cost recurs every month and falls on households whose income the illness has already reduced.

A financing system in which the state pays US$7.74 per person can deliver a vaccination campaign. It cannot deliver forty years of antihypertensives.

What would change this conclusion

A specific, checkable test: if the government's share of current health expenditure rises above 25% and the out-of-pocket share falls below 70% by 2030, the argument here is wrong, and Bangladesh will have restructured its financing before the disease transition forced it to. If the out-of-pocket share is still near 79% in 2030, the transition will be arriving against an unchanged financing model.

Three moves, with owners and signals.

1. Put a floor under public health financing in absolute per-person terms, not as a share of a growing budget. Owner: Ministry of Finance, in the FY2027-28 budget cycle. Success signal: domestic general government health expenditure rising above 0.5% of GDP in the next WDI release, which would still leave Bangladesh last of six but would end two decades of drift.
2. Buy the recurring costs first, not the buildings. A drug benefit covering hypertension and diabetes medicines at upazila level attacks the exact category that out-of-pocket financing handles worst, and it costs a fraction of hospital construction. Owner: Health Services Division. Success signal: out-of-pocket share below 75% in the WDI series, the first sustained fall in over a decade.
3. Report catastrophic health expenditure as a standing national statistic. Bangladesh manages what it measures, and it does not currently measure, on any regular published basis, how many households are pushed under the poverty line by medical bills. Owner: BBS, alongside HIES. Success signal: a published incidence series, which would make the cost of the current model visible in a way that a share-of-GDP figure never will.

The counterargument

The strongest objection is straightforward: Bangladesh delivered the region's longest male life expectancy on the region's smallest public health budget, so why treat the financing model as a failure rather than an efficiency to be admired and copied.

Two answers. First, the model's cheapness and its results are less connected than the pairing suggests. Fertility decline, female education, rural electrification and sanitation did much of this work, and they sit outside the health budget entirely. Crediting the financing structure for outcomes produced largely beside it is how a country talks itself into keeping a structure past its usefulness.

Second, the objection measures the wrong thing. A system that produces good average survival while charging households US$42.40 a head at the moment of illness is not efficient; it has moved the cost somewhere the national accounts do not show it, onto the balance sheets of families. The efficiency is an accounting artefact. Bangladesh's health system is not inexpensive. It is expensive, and the bill is simply not addressed to the state.

---

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 comparators reporting in the stated year, excluding Bangladesh. Health financing figures are 2023, life expectancy 2024, physicians and hospital beds 2021, being the latest years with comparable coverage.

Sources

1. World Bank WDI, Domestic general government health expenditure (% of GDP), SH.XPD.GHED.GD.ZS: https://data.worldbank.org/indicator/SH.XPD.GHED.GD.ZS
2. World Bank WDI, Current health expenditure (% of GDP), SH.XPD.CHEX.GD.ZS: https://data.worldbank.org/indicator/SH.XPD.CHEX.GD.ZS
3. World Bank WDI, Out-of-pocket expenditure (% of current health expenditure), SH.XPD.OOPC.CH.ZS: https://data.worldbank.org/indicator/SH.XPD.OOPC.CH.ZS
4. World Bank WDI, Current health expenditure per capita (current US$), SH.XPD.CHEX.PC.CD: https://data.worldbank.org/indicator/SH.XPD.CHEX.PC.CD
5. World Bank WDI, Life expectancy at birth, male (years), SP.DYN.LE00.MA.IN: https://data.worldbank.org/indicator/SP.DYN.LE00.MA.IN
6. World Bank WDI, Life expectancy at birth, female (years), SP.DYN.LE00.FE.IN: https://data.worldbank.org/indicator/SP.DYN.LE00.FE.IN
7. World Bank WDI, Life expectancy at birth, total (years), SP.DYN.LE00.IN: https://data.worldbank.org/indicator/SP.DYN.LE00.IN
8. World Bank WDI, Physicians (per 1,000 people), SH.MED.PHYS.ZS: https://data.worldbank.org/indicator/SH.MED.PHYS.ZS
9. World Bank WDI, Hospital beds (per 1,000 people), SH.MED.BEDS.ZS: https://data.worldbank.org/indicator/SH.MED.BEDS.ZS