Executive finding
Pre-primary enrolment is 36.38% gross, roughly half the peer median of 74 and the lowest of the four Asian economies reporting. Bangladesh clears the primary school gate above the peer median and falls behind on both sides of it.
Executive Summary. Bangladesh is above its peer median at exactly one point in the schooling sequence: the moment children enter primary school. Primary enrolment was 106.5% gross in 2024, second of six Asian economies and 5.487 points above the peer median of 101. On either side of that gate the position reverses. Pre-primary enrolment was 36.38% gross in 2021, the lowest of the four countries reporting and 49.2% of the peer median of 74, a shortfall of 37.62 percentage points. Downstream, primary completion was 94.3% in 2024, 3.236 points under the median of the five reporting, and lower secondary completion was 73.81%, fifth of six and 21.4 points under the peer median of 95.21. Read in order: a deficit of 37.62 points before school, a surplus of 5.487 at entry, deficits of 3.236 and 21.4 on the way out. Public education spending, 2.16% of GDP in 2023, was second of the four countries reporting and above their median of 1.947, and at 2.034% in 2024 it is a smaller share of output than the 2.125% of 2000. The question this data poses is less how much the state spends on schooling than at which point in a child's life it shows up.
The year that mostly does not happen
Pre-primary enrolment in Bangladesh was 36.38% gross in 2021, the lowest of the four countries reporting that year. Vietnam recorded 92.45%, Thailand 74%, India 51.69%. Pakistan and Indonesia have no 2021 observation in this series, so the comparison rests on four countries, not six, and nothing here supports a claim about the region as a whole.
Against the peer median of 74%, Bangladesh sits 37.62 percentage points short and reaches 49.2% of it. Roughly half.
Source: World Bank World Development Indicators, school enrolment, preprimary (% gross, SE.PRE.ENRR), 2021.
Three qualifications, and they matter more here than usual.
First, this is a gross enrolment ratio: enrolment of any age over the population of the official pre-primary age. It measures places filled, not children of the right age covered, and it moves when a country changes how many years of pre-primary it defines. A system with one pre-primary year and a system with three are not producing the same denominator.
Second, 2021 is the latest year all four report, and it is a year in which early-childhood provision was widely suspended and slowly restored. Part of the Bangladeshi reading is closure rather than structure, and this data cannot separate the two.
Third, cutting the other way: that year applied to everyone. Vietnam recorded 92.45% in the same pandemic year. Whatever 2021 did to attendance, it did not open a 37.62 point gap on its own.
The mechanism worth naming is not a research finding, it is arithmetic about what a school day holds. A child who reaches grade one without a structured year behind them starts further back than a classmate who had one, and a system carrying 30.05 pupils per teacher leaves little room to close that distance child by child. The deficit does not stay in grade one. It compounds through every grade that assumes the previous one worked.
The gate is the strong number
Primary enrolment is the one stage of the sequence where Bangladesh sits above its peers, and the figure is genuinely good. At 106.5% gross in 2024, Bangladesh was second of the six, behind India's 120.5% and ahead of Vietnam's 106.3%, Thailand's 101%, Indonesia's 100.0% and Pakistan's 83.73%. The peer median is 101%, so Bangladesh runs 5.487 points above it, at 1.054 times the median.
A gross ratio above 100% does not mean more children are enrolled than exist. It counts pupils of any age against the official primary-age population, so it rises with over-age and under-age entrants, which is what late entry and grade repetition produce. A ratio inflated by late entrants is partly a symptom of the missing year before school, not independent evidence against it.
The reading is narrow and still important: getting children through the primary school door is not the constraint, and nothing in this data suggests reopening that question.
The losses are downstream
Follow the same sequence to the exits and the position reverses.
Primary completion was 94.3% in 2024, fourth of the five countries reporting (Indonesia has no 2024 observation in this series). India recorded 94.49%, Vietnam 100.6%, Thailand 102.6%, Pakistan 74.14%. Bangladesh is 3.236 points below the peer median of 97.54%, at 0.9668 of it. On its own that is a modest gap.
Lower secondary completion is where the gap opens. Bangladesh recorded 73.81% in 2024, fifth of the six, ahead of Pakistan's 55.82% and behind India's 87.28%, Vietnam's 95.21%, Indonesia's 101.7% and Thailand's 102.8%. The peer median is 95.21%, which puts Bangladesh 21.4 points short, at 77.5% of the median.
Source: World Bank WDI, lower secondary completion rate (% of relevant age group, SE.SEC.CMPT.LO.ZS), 2024.
Set the four positions in order, all from 2024 except the first:
- Pre-primary enrolment (2021): 37.62 points below the peer median.
- Primary enrolment: 5.487 points above the peer median.
- Primary completion: 3.236 points below.
- Lower secondary completion: 21.4 points below.
The advantage exists at the moment of entry and is gone by the end of primary. By the end of lower secondary the shortfall is the largest anywhere in the sequence apart from the pre-primary year itself. One caution: completion rates are measured against the relevant age group, not against an entering cohort, so these four indicators are separate constructions rather than one set of children followed through. They establish a pattern of position, not a traced pathway.
What the classroom holds
Two indicators describe the room where the loss happens. Both have problems worth saying out loud.
Bangladesh had 30.05 pupils per primary teacher in 2018, second of the five countries reporting that year, behind Pakistan's 44.08 (India has no 2018 observation). Indonesia recorded 17.03, Vietnam 20.28, Thailand 16.64. The peer median is 18.66, so the Bangladeshi classroom carried 11.4 more pupils per teacher, 1.611 times the median load.
That figure is six years older than the completion data beside it and may have improved since. It is also a national average, and the average is not what determines whether a child is taught: a mean of 30.05 is compatible with far heavier classes in the schools serving the children most likely to leave early.
On teacher quality, 77.88% of Bangladeshi primary teachers were trained in 2024, fourth of the six, 11.37 points below the peer median of 89.25% and at 0.8726 of it. India recorded 91.73%, Thailand 100%, Vietnam 89.25%, Pakistan 74.97%, Indonesia 33.51%.
Indonesia is the reason not to lean on this indicator. A country reporting 33.51% trained teachers also reports lower secondary completion of 101.7%, which is not a combination any causal story accommodates. "Trained" is a nationally defined category, and the definitions are not comparable across these six countries in any strict sense. The Bangladeshi figure is a fact about Bangladeshi certification rules at least as much as about classroom competence.
The budget answer is thinner than it looks
The intuitive explanation is that Bangladesh underspends. This data neither confirms that nor clears it, and the reason is worth spelling out.
Public education spending was 2.16% of GDP in 2023, second of the four countries reporting, behind Thailand's 2.521% and above Pakistan's 1.947% and Indonesia's 1.275%. The peer median of those three is 1.947%, so Bangladesh sits 0.216 points above it, at 1.111 times the median.
Treat that as thin. India and Vietnam have no 2023 observation in this series, and both complete lower secondary at higher rates than Bangladesh (87.28% and 95.21%). A median built from three countries is not a regional benchmark, and this essay does not claim Bangladesh outspends its region. The chart makes the limitation visible: in every year since 2000 that India and Vietnam do report, both sit above Bangladesh. Second of four reporters in 2023 describes who filed a return that year, not generosity.
The reading that survives is the Bangladeshi line on its own.
Source: World Bank WDI, government expenditure on education (% of GDP, SE.XPD.TOTL.GD.ZS), Bangladesh against comparators, 2000-2024.
Public education spending was 2.125% of GDP in 2000 and 2.034% in 2024: down 0.09 points, 4.267% lower, 0.957 times its level at the start of the century. That is not a story about schools receiving less cash. A steady share of a growing economy is more money, and the economy grew (GDP was US$450.1 billion in 2024, expanding at 4.223% that year, in a country of 173.6 million people). It is a story about the claim education holds on national income, which is no larger than it was in 2000.
A share of GDP is also not spending per pupil, and this data contains no per-pupil series. A country with a young age structure buys fewer teacher-hours per child out of the same share of output than an older one does, which is a reason the flat share may understate the squeeze rather than overstate it. A budget frozen as a share of output is a budget that can only add a new stage of schooling by taking from an existing one. The pre-primary year is exactly the kind of addition that loses that argument every year.
What would change this conclusion
A specific, checkable test. If the next vintage of this series shows Bangladeshi pre-primary enrolment above India's 51.69%, and the 21.4 point lower secondary completion gap against the peer median narrows across two consecutive releases, the argument here is wrong and the 2021 reading was pandemic damage now repaired. If pre-primary is still near 36.38% and lower secondary completion still near 73.81% in the 2035 data, then a decade will have been spent enrolling children into a sequence that keeps losing them at the same point.
Three moves, with owners and signals.
- Fund the pre-primary year as a budget line with its own teachers, not as an unpaid extension of grade one. The gap to close is 37.62 percentage points, the largest anywhere in this data. Owner: Ministry of Primary and Mass Education, in the next annual development programme cycle. Success signal: pre-primary gross enrolment above India's 51.69% in the first WDI release reporting Bangladesh after 2021.
- Treat the primary-to-secondary transition as the failure point, because that is where the gap multiplies. The shortfall against the peer median widens from 3.236 points at the end of primary to 21.4 points at the end of lower secondary, so interventions aimed only at primary retention address the smaller half of the problem. Owner: Ministry of Education, secondary directorate. Success signal: lower secondary completion above 87.28%, India's current level, before the pre-primary cohorts now missing reach that grade.
- Restore reporting on classroom inputs. The most recent comparable pupil-teacher observation in this series is 2018, and the trained-teacher share is a nationally defined category with no cross-country meaning. Owner: BANBEIS, in the annual education statistics release. Success signal: a post-2018 pupil-teacher figure in the WDI series, and a trained-teacher share published against a definition an outside analyst can replicate.
The counterargument
The strongest objection attacks the lead indicator directly, and it lands.
Pre-primary gross enrolment is the weakest series in this pack. The reading is from 2021, a year of suspended provision. It is a gross ratio, sensitive to how many pre-primary years a country defines. Only four of the six countries report it. And to the extent that a formal pre-primary year is a recent addition to the state system, a low ratio partly measures the age of the policy rather than a refusal to fund it. On this reading, the essay has taken the noisiest number available and hung a sequence on it.
Three answers, the last of which is a concession.
First, the argument does not rest on the pre-primary figure alone. Lower secondary completion is a separate series, a separate year, and a separate construction, and it puts Bangladesh 21.4 points below the peer median with only Pakistan lower. Remove pre-primary entirely and the downstream loss is still there, still large, and still unexplained by the enrolment figure that precedes it.
Second, the objection about the policy's age is a reason to act, not to discount. A gap that exists because the state began funding the year late is still a gap in the cohort entering grade one now, and that cohort does not get a second chance at being five.
Third, what this cannot settle. There is no learning-outcome data in this comparison: no assessment scores, no learning-adjusted years of schooling, nothing measuring what a completed grade contains. Completion is attendance, not learning, and a system could close every gap listed here without a child reading better. Six countries are a comparison, not a sample, and every association in this essay is cross-sectional. Vietnam's 92.45% pre-primary enrolment and 95.21% lower secondary completion sit beside each other in the data; nothing here shows the first produced the second. Confounders run through all of it: income levels, the age structure of the population, how far children travel to school, how early they are needed at work. The pattern is real and the ordering is consistent. The causal claim is not made.
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. Years differ by indicator because coverage does: enrolment, completion, trained teachers and the spending trend are 2024, cross-country spending 2023, pre-primary enrolment 2021, the pupil-teacher ratio 2018, each the latest year with comparable coverage. Pre-primary enrolment and cross-country spending report four countries, primary completion and the pupil-teacher ratio five; every comparison drawn from them is stated on that basis.
Sources
- World Bank WDI, School enrollment, preprimary (% gross), SE.PRE.ENRR: https://data.worldbank.org/indicator/SE.PRE.ENRR
- World Bank WDI, School enrollment, primary (% gross), SE.PRM.ENRR: https://data.worldbank.org/indicator/SE.PRM.ENRR
- World Bank WDI, Primary completion rate (%), SE.PRM.CMPT.ZS: https://data.worldbank.org/indicator/SE.PRM.CMPT.ZS
- World Bank WDI, Lower secondary completion rate (%), SE.SEC.CMPT.LO.ZS: https://data.worldbank.org/indicator/SE.SEC.CMPT.LO.ZS
- World Bank WDI, Pupil-teacher ratio, primary, SE.PRM.ENRL.TC.ZS: https://data.worldbank.org/indicator/SE.PRM.ENRL.TC.ZS
- World Bank WDI, Trained teachers in primary education (%), SE.PRM.TCAQ.ZS: https://data.worldbank.org/indicator/SE.PRM.TCAQ.ZS
- World Bank WDI, Government expenditure on education (% of GDP), SE.XPD.TOTL.GD.ZS: https://data.worldbank.org/indicator/SE.XPD.TOTL.GD.ZS
Cite this
BDPolicyLab Research. (2026). School Starts Too Late: The Missing Year Before Grade One. BDPolicyLab. https://bdpolicylab.com/publications/school-starts-too-late-the-missing-year-before-grade-one
Method and source
Source: Primary sources cited at point of use in the publicationAs of 10 Aug 2026