OUT-OF-SCHOOL CHILDREN IN SINDH: Peeling back, layer by layer, every explanation we give ourselves for why almost half of Sindh's children are not in school

 

Every explanation in this article is tested against the same evidence: the 7th Population and Housing Census 2023, the Annual School Census series of 2013-14, 2018-19 and 2024-25, and the district-wise school-specific budget of the School Education and Literacy Department. Some of the explanations survive. Most do not.

The number, and what it means

According to Census 2023, there are 16,891,397 children aged five to sixteen in Sindh. Of them, 7,818,248 are not in any school. That is 46.3% — almost one child in every two.[1][2]

In plain words: if you see a hundred children inside a school in Sindh, very nearly the same number are standing outside it. Girls are 52.3 % of them.

What does this mean for Sindh's future? A child who is eight years old today and has never been enrolled will be eighteen in 2036 and thirty in 2048. Sindh is manufacturing the unqualified, unskilled, and least-paid workforce of 2050 right now. The province added 7.8 million people between the 2017 and 2023 censuses. The cohort is getting larger, not smaller.[3]

Sindh is not late in educating these children. It is, at current rates, not going to educate most of them at all.

The legal position is not ambiguous either. The Sindh Right of Children to Free and Compulsory Education Act, 2013 makes schooling a justiciable entitlement for exactly this age group, five to sixteen.[4] The 7.8 million are children the state is legally obliged to have reached, but failed to.

So why are they out of school? Let us take the explanations one at a time, starting with the simplest, and let us make each one answer to the data.

Figure 1. Out-of-school children as a share of all children aged 5-16, Census 2023. Thatta, Sujawal and Kashmore lose roughly three children in four.

District	Children 5-16	Out of school	Out of school %	Never attended %	Female literacy %	Facility deprivation %	Women as % of teachers
Thatta	354,248	268,486	76	94	20.1	51	31.0
Sujawal	277,334	207,764	75	95	19.7	62	20.4
Kashmore	434,870	304,236	70	95	25.7	52	16.6
Shikarpur	487,742	307,001	63	90	34.6	34	24.6
Tando Muhammad Khan	226,997	141,821	62	93	25.0	34	27.4
Jacobabad	414,790	257,173	62	91	32.6	39	24.7
Badin	622,448	381,384	61	92	26.0	46	25.8
Tharparkar	611,420	372,976	61	93	23.5	49	11.8
Tando Allahyar	287,015	170,489	59	90	31.5	24	27.7
Sanghar	739,844	418,695	57	91	33.9	36	27.9

Table 1. The ten districts with the highest out-of-school rates. Out-of-school children aged 5-16 and female literacy: Census 2023. Facility deprivation is the average share of government schools lacking electricity, a washroom, drinking water or a boundary wall; female teaching share: Annual School Census 2024-25.

 

Layer one: “There are not enough schools”

This is the first thing almost everyone says, and it is the easiest to test. If the shortage of schools were the binding constraint, districts with more schools per child would have fewer children out of school.

Sindh has 41,157 government schools for 16.89 million children — about 24 schools for every 10,000 children of school age. But the relationship between school supply and out-of-school children runs the wrong way. Districts with more schools per child tend to have more children out of school, not fewer (r = +0.67).

Figure 2. If school supply were the constraint, this line would slope downwards. It does not.

The clearest illustration is Karachi. Karachi East has 1.7 government schools per 10,000 children — the thinnest provision in Sindh — and 27% of its children out of school. Tharparkar has more than ten times as many schools per child and 61 % out of school. Across the seven Karachi districts, 32.1% of children are out of school; across the rest of Sindh, 53.3%.

Let me be honest about the limits of this test, because the counter-argument is a fair one. Karachi does much of the statistical work here: remove the seven Karachi districts and the correlation falls to r = +0.29. Karachi's schooling is overwhelmingly private — 91.9% of its primary enrolment sits in private schools — so its low government-school count and its low out-of-school rate have a common cause rather than a causal link.

But that concession does not rescue the argument. Even confined to rural Sindh, more schools per child never predicts fewer children out of school. The sign never turns negative.

There is a second reason to doubt the supply story. Of Sindh's 41,157 schools, 2,482 — 6% — were recorded as closed in 2024-25. In Jamshoro, 14.8% of schools are shut, in Sujawal 11.8%, in Umerkot 11.4%. A province that cannot keep the schools it has open does not have a shortage of buildings.

The lack of schools does not explain why 46 % of Sindh's children are out of school. So what does?

 

 

Layer two: “Children drop out after primary”

This is the most respectable version of the argument, and it starts from a real structural fact. Sindh's school network is overwhelmingly primary: 36,356 of 41,157 government schools — 88% — are primary schools. Middle, elementary and secondary schools together are just 10.5% of the network, and they cluster in towns. For a girl in a village, getting to class six can mean leaving the village.

The enrolment staircase seems to confirm it. Of every 100 children in class I in 2024-25, 43 are in class VI, 40 in class VIII, 36 in class X, and 14 in class XII.

Figure 3. Of every 100 children enrolled in class I, 63 remain by class V — the largest single loss happens before primary school ends. Annual School Census 2024-25.

But look more carefully at where the staircase actually breaks. Of every 100 children in class I, only 63 are still there in class V. Thirty-seven of them are gone before primary school even ends — before the shortage of middle schools can possibly be the cause.

Children in Sindh do not mainly leave school when the schools run out. They leave while the schools are still there.

One methodological caution, since this figure is widely quoted. These are cross-sectional ratios: they compare different children in different classes in the same year, not one cohort followed over twelve years. Because enrolment has been growing, they somewhat overstate true dropout. The qualitative conclusion is unaffected — the steepest single fall, from class V to class VI, is a 32 % drop, but the cumulative loss inside primary is larger.

And now the decisive objection. Dropouts — children who attended school and then left — are 1,201,900, or 15.4% of all out-of-school children. The other 6,616,348, fully 84.6%, never attended a single day.

 

Figure 4. The composition of Sindh's out-of-school children. Retention policy addresses the gold segment.

Read that again, because it reorganises the entire problem. Six of every seven out-of-school children in Sindh were never enrolled. Retention policy, transition rates, middle-school upgrades and dropout campaigns all address the smaller sixth.

 

Layer three: “They are working”

If most out-of-school children never enrolled, perhaps they were never available to enrol. Poor households need their children's labour; the child who herds goats or works a loom cannot sit in a classroom. This is the most humane version of the argument, and it feels true.

The census does not support it. Of Sindh's 16.89 million children aged five to sixteen, 1,462,733 are recorded as employed — 8.7%. Even if we look only at the children who are out of school, just 18.7% of them are working.

 

Figure 5. Child labour explains a fifth of the problem. It leaves 6.36 million children unaccounted for.

 

Four out of every five out-of-school children in Sindh — 6.36 million of them — are neither in school nor in work. They are simply idle.

That is the single most uncomfortable number in this article. These children are not being kept away by an economic necessity that a stipend or a cash transfer could offset. Nothing is competing for their time.

Two honest qualifications. First, Census 2023 records zero employment for every child aged five to nine, so measured child work effectively begins at ten; labour among the youngest is not captured, and the true work rate is higher than 8.7%. Second, unpaid domestic work — overwhelmingly done by girls — is not counted as employment at all. Both caveats matter, and neither is large enough to close a gap between 8.7 % working and 46.3% out of school.

Layer four: “The schools are simply too far away”

Distance is the explanation most often offered for girls specifically, and it is easy to test using district area against school count.

It does not carry much weight. The area served per government school correlates with the out-of-school rate at only r = +0.34. Jamshoro has the most thinly spread network in Sindh — 12.7 square kilometres per school — and an out-of-school rate of 49%, well below Thatta's 76% or Sujawal's 75%, both of which have roughly half Jamshoro's area per school.

Physical distance is real for individual families, particularly beyond primary. But as a district-level explanation for a 46% out-of-school rate, geography is a weak predictor. Sindh's problem is not principally that its children live too far from schools.

Layer five: “The schools are there, but there are not enough teachers”

A school without a teacher is no school. We know there are hundreds of single-teacher schools in Sindh, but the annual school census does not provide a district-wise breakdown of schools by the number of teachers, but mentions only a cumulative figure of schools and teachers in each district. On average, Sindh has 3.9 teachers per school – Karachi South has 18.2, Hyderabad has 9.8, and Tharparkar has 1.8. Despite that, the data does support the argument that lack of teachers is negatively correlated with the number of OOSC, i.e., districts with fewer teachers have a higher percentage of OOSC.

 

Figure 6. Shortage of teachers is negatively correlated to the percentage of OOSC

But when it comes to the education of girls, what determines whether a family sends a daughter to school? The most probable answer is the presence of a female teacher.

The data supports it strongly. Across the thirty districts, the percentage of female teachers correlates with the girls’ out-of-school rate at r = -0.87. Women are 33.3% of Sindh's government teachers overall, but 11.8% in Tharparkar, 16.6% in Kashmore and 20.4% in Sujawal — against 73.1% in Karachi Korangi. The districts with the fewest women teaching are the districts with the most girls out of school.

 

 

Figure 7. More female teachers inside a school, fewer girls out of school.

 

But we must be cautious here: correlation is not causation, and here the causal arrow could easily run backwards. A district with few educated women will have few women qualified to teach. The absence of female teachers might be a consequence of the same history that produces out-of-school girls, not its cause.

 

Layer six: “The schools are there, but they are not fit to attend”

If a school has no boundary wall, no usable toilet, no drinking water and no electricity, a parent deciding whether to send an eight-year-old daughter may reasonably decide not to.

Taking the average share of schools in a district lacking each of four basic facilities — electricity, a washroom, drinking water and a boundary wall — facility deprivation correlates with the out-of-school rate at r = +0.91. It is one of the strongest relationships in the entire dataset.

Figure 8. Facility deprivation is one of the two strongest correlates of out-of-school children in Sindh.

The range is enormous. In Sujawal, 62% of schools lack the average basic facility; in Karachi South, 2% do. Sujawal has 75% of its children out of school; Karachi South has 26%.

Province-wide, the picture has improved unevenly since 2013-14. Schools with drinking water rose from 50.7 to 74.4% and schools with a washroom from 56.7 to 77.1% — real gains. But schools with electricity stood at 50.3% in 2013-14, fell to 35.1% in 2018-19, and are still only 47.9% today, below where they started.[5][6][7]

So we appear to have found a reason that the data supports. Before we accept it, one caution: facility deprivation is also strongly correlated with everything else that is wrong in a district. We will return to the question of what is cause and what is company.

 

Layer seven: “It is gender norms — and that is why we pay a stipend”

Girls are 52.3% of out-of-school children, so any serious account must deal with gender. Sindh's flagship instrument is the Girls' Stipend Programme, and it is worth examining closely, because its design tells us what the state believes the problem is.

The facts, from the Department's own account.[8] The stipend has been distributed for more than two decades. It covers girls in classes VI to X. Two rates were set following a World Bank survey in 2006: Rs 3,500 a year in 45 talukas under a Differential Stipend Policy, and Rs 2,500 in the rest. Those rates have now been equalized at Rs 4,500. For the 2025-26 academic year, eligibility was narrowed further, to girls in classes VI, IX and X with at least 60 % attendance.[9]

Three things follow, and each of them is a problem.

First, the amount. Rs 4,500 a year is under Rs 375 a month, or roughly Rs 12.5 a day. In the five years to the end of 2025 alone, cumulative inflation in Pakistan was about 100% — prices doubled.[10] A stipend has therefore lost roughly half its real value since 2020, and far more since 2006. To continue her education beyond primary school, a girl has to move out of her village. Do you think a rickshaw or a van would pick up and drop her for Rs 12.5 a day?

Second, the reach. The Department's own published series shows beneficiaries falling from 344,744 in 2013-14 to 278,000 in 2016-17 and 260,000 in 2020-21, even as the school-age population grew.[8]

Third, and most fundamentally: the stipend is paid to girls who are already in class VI. It is an instrument for retaining girls who have already been enrolled, in a province where 85 % of out-of-school children were never enrolled at all.

The census makes this mismatch precise. Girls are 53% of children who never attended school, but only 48.3 % of dropouts. In other words, once a girl in Sindh is enrolled, she is very slightly less likely than a boy to leave. The entire female disadvantage is concentrated at the front door.

This is not an argument against the stipend, which does useful work for the girls who receive it. It is an argument that Sindh's principal gender instrument is aimed at the wrong end of the pipeline, in an amount that inflation has hollowed out, reaching fewer girls each year.

  

Figure 9. The Girls' Stipend has been Rs 3,500 since it was equalised, while beneficiaries fell and prices doubled in the five years to 2025 alone.

Layer eight: so is it the mothers' generation?

The way to test that is to look past the teaching cadre to the stock of female education in the district as a whole. If female literacy predicts out-of-school children even better than the female teaching share does, then teacher supply is largely a symptom.

It does. Female literacy correlates with the district out-of-school rate at r = -0.93 — the strongest single relationship in this analysis, stronger than facility deprivation, stronger than the teaching cadre, far stronger than schools per child or budget.

Figure 10. The strongest single relationship in the data. Today's out-of-school children are the children of yesterday's out-of-school girls.

The best available predictor of whether a child in Sindh is in school today is whether the women of that district were in school twenty years ago.

This is what makes the problem self-reproducing. A district without literate women cannot supply female teachers; without female teachers it enrols fewer girls; those girls become the next generation of illiterate women. Sujawal, with 19.7% female literacy, has 75% of its children out of school. Karachi Central, with 83.3% female literacy, has 23%.

The practical implication is uncomfortable but clear. A loop cannot be broken from inside itself. Recruiting female teachers locally in Tharparkar cannot work when the local pool of educated women does not exist. Breaking it requires importing teachers on residential terms, creating a bridging cadre with in-service qualification, or building the pool deliberately a decade ahead — and probably all three.

Layer nine: “Then it must be money”

The final reflex is fiscal: that the districts with more schools deprived of basic facilities are districts that receive less. The school-specific budget amounting to PKR 18.67 billion has recently been allocated for 88.2% of schools in the province. Ironically, the allocation of the budget bears no relationship with the deprivation of facilities. Urban districts of Karachi and Hyderabad, where schools are the least deprived of basic facilities, receive the largest share of funds, whereas schools in rural and remote areas, where school deprivation is high (and where the number of OOSC is high too), receive less funding. It is understandable that urban schools require more funds for operation, maintenance, and utility costs, and their funds should not be reduced. But the schools, most deprived of basic facilities, should be allocated more funds.

Figure 11. The most deprived schools received the least budget

 

But the story does not end here. Across the province, only 52.6% of the school-specific budget had been released, and the release rate correlates with the out-of-school rate at r = -0.48: districts with more children out of school get a smaller share of their own allocation. Tharparkar had 45.8% released, Sujawal 45.4%, against 53 to 57% in better-off districts.

There is a scale point too. The school-specific budget — the only money a head teacher can actually spend on the school — totals Rs 18.67 billion, about 4.2% of the School Education Department's Rs 446.96 billion for 2026-27.[12] Roughly 96% of the school education budget is salaries and establishment. The province can pay a teacher to stand in a room it cannot afford to light.

Layer ten: “But we are now monitoring the schools”

Sindh, it is said, has fixed the accountability problem. The School Education and Literacy Department now runs the Sindh School Monitoring System, which collects attendance and school-condition data every month and pushes it to a web dashboard that officials can read in something close to real time.[17] Teachers mark themselves present on portable biometric devices. A Student Attendance Monitoring and Redress System tracks pupils.[18] In August 2026 the province approved FRAMES — a facial-recognition attendance system, geo-tagged to each school and able to run offline — with a pilot in one district from each of Sindh’s six divisions before provincial rollout.[19] No other province has gone this far.

So let us ask the only question that matters. Has any of it improved what children learn?

First, let us give the monitoring its due, because it has achieved something real. The Asian Development Bank’s 2023 review of school teaching in Pakistan found that teacher absenteeism did decline in both Punjab and Sindh after biometric monitoring and unscheduled inspections were introduced.[20] ASER’s independent enumerators, who arrive unannounced, recorded teacher attendance of 89% in rural schools and pupil presence of 75% in government schools in 2023, the latter up on 2021.[15] Bodies are in rooms. That is not nothing, and anyone who has worked in a district where a third of the staff never appeared will tell you it is a great deal.

Now the learning. ASER tests children on grade-two competencies — the floor, not the ceiling. In rural Sindh, of children enrolled in class V, the share who could read a class-two story in Urdu or Sindhi went from 40% in 2021 to 39% in 2023. English sentence reading went from 23% to 22%. Two-digit division went from 28% to 27%.[15]

Every learning indicator moved in the wrong direction, by roughly a point, across exactly the years in which the monitoring architecture was being built.

And the same comparison exposes the ceiling of the presence story. ASER recorded rural teacher attendance in Sindh at 89% in 2019 as well.[21] Four years, one biometric rollout and one monitoring dashboard later, the independently measured figure is the same 89%. The gains the ADB describes were real, but they were largely banked before the current systems arrived, and the metric has now flattened at a level that still means one teacher in nine is absent on any given day.

Look one grade lower and the floor gives way completely. In class three, ASER 2023 found that 8% of children in Sindh could read an Urdu story, 5% could read English sentences, and 3% could do a two-digit division — all tasks written for class two.[15] These are children who are enrolled, whose teachers are marked present, and who are themselves recorded present. Every indicator the dashboard collects is green. But the children cannot read.

Let me put the whole chain in one line, because it is the most important sentence in this article. Take 100 children of school age in Sindh. Fifty-four are enrolled anywhere at all. Of those, on current class-wise ratios, about 34 are still enrolled by class V. Of those, on ASER’s rural results, about 13 can read a story written for class two.

Sindh does not have 7.8 million children outside the system and 9.1 million inside it. It has 7.8 million outside, and roughly 13 in every 100 who pass through the middle of the system with the skill the system exists to impart.

I owe you three honest caveats on that cascade. The 54 and the 34 come from province-wide census and school-census data covering all areas; the 13 rests on ASER’s rural sample, and rural Sindh learns less than urban Sindh, so the provincial figure would be somewhat higher. The class-V ratio is cross-sectional, as noted in Layer two. And the three steps measure three different groups of children, not one cohort followed through. The cascade is an illustration of magnitude, not an estimate. But no defensible adjustment to any of the three steps rescues the conclusion.

Why would monitoring not work? Because it measures the one input that was never the binding constraint on learning. Consider what the dashboard cannot see. Sindh has 159,352 government teachers across 41,157 schools — 3.9 teachers per school on average, and 36,356 of those schools are primary schools with five or more grades to cover. In Tharparkar, there are 1.8 teachers per school. In Umerkot 2.3, in Sujawal 2.3, in Badin 2.6, in Thatta 2.9. In Karachi South, there are 18.2. ASER’s own conclusion on Sindh is blunt: it ranks lowest in foundational learning nationally, with the majority of primary schools being one- or two-teacher schools.[15]

That number turns out to matter more than the number of schools. Teachers per school correlates with the district out-of-school rate at r = −0.80 — nearly as strong as female literacy, and in the direction common sense expects. Recall from Layer one that schools per 10,000 children ran the wrong way, at +0.67. Sindh’s problem was never a shortage of buildings. It is that the buildings were built thin.

A biometric device confirms that the one teacher is present. It does not produce the second teacher.

Nor can monitoring supply what a teacher needs to teach. In 2019, of 133,076 teachers in Sindh, exactly 581 were subject specialists — 0.4 % — and only 189 of those were women.[20] A monitoring system tells you that a person unqualified to teach mathematics arrived on time, not to teach it.

The international evidence points the same way, and it is worth hearing because it comes from the province next door. When Punjab piloted performance pay for primary teachers, the evaluation found that the bonus raised enrolment and examination participation — but not examination scores.[20] Attendance-shaped incentives buy attendance. They do not buy learning. The ADB’s own recommendations, tellingly, move past monitoring altogether: hire teachers locally, create advancement paths inside teaching rather than out of it, fix pre-service preparation, and close the subject-specialist gap.[20]

And here is where this layer closes the circle back to the 7.8 million. Remember that 84.6% of Sindh’s out-of-school children never enrolled at all. Enrolment is a household decision, and households are not blind. A mother in Tharparkar who watches a neighbour’s son attend a two-teacher school for five years and emerge unable to read a paragraph has been given evidence, not a prejudice. The dashboard glows green in Karachi while she makes her decision on what she can see with her own eyes.

The monitoring reforms are aimed at the 15% who leave. The learning results are the reason the 85% never come.

None of this is an argument against monitoring, and I want to be precise about that, because the argument is easily misused. Attendance data is a genuine public good; a state that cannot see its own schools cannot govern them, and Sindh can now see them. FRAMES may well push teacher attendance above 89%, and it should. The error is not in building the system. The error is in reading presence as progress, and in the fact that after more than a decade of monitoring investment Sindh still has no published, district-wise, comparable measure of whether children can read. We monitor the input we can automate, and we do not measure the output we exist to produce.

If a single recommendation follows from this layer, it is this: extend the monitoring architecture that already reaches every school to a sample-based, district-wise foundational learning assessment, published annually alongside the Annual School Census. Sindh has built the pipes. It is sending the wrong water through them.

 

What the evidence actually says

Put every explanation on the same axis, and the ranking is unambiguous. The variables that move most closely with out-of-school children in Sindh are female literacy (r = -0.93), facility deprivation (r = +0.91), urbanisation (r = -0.84), and the female teaching share (r = -0.82). School supply per child (r = +0.67, and only +0.29 outside Karachi) has the wrong sign. Most deprived schools receive the least funding.

Four conclusions follow, and they are not the ones our policy documents are built around.

One. This is an enrolment crisis, not a retention crisis. 85% of these children never came. Every instrument designed around keeping children in school addresses a sixth of the problem.

Two. It is not principally about buildings, distance or child labour. Sindh has schools its children do not enter, and 6.36 million children who are neither in school nor working.

Three. It is about whether a school is worth entering — facilities, and above all a female teacher — and about whether the household has any history of education at all.

Four. The gender problem is at the front door, not the back. Girls are 53% of the never-enrolled and 48% of dropouts. Sindh's flagship girls' instrument pays from class VI.

If the province did only one thing differently, it would be this: move the money, the campaigns and the incentives from retention to first enrolment, and concentrate them in the districts where women cannot read.

None of this is a counsel of despair. The same data that indicts the current approach also identifies precisely where to act: eight districts below 25% female teaching staff, a facilities gap that is measurable school by school, a release rate that can be fixed one financial year, and 6.36 million children whose time nothing else is currently claiming. The children are, for now, still there to be enrolled. In ten years they will not be children, and the question will no longer be one for the education department.

References

[1]  Pakistan Bureau of Statistics. 7th Population and Housing Census 2023: Sindh district tables — Table 4, population by age; Table 12, literacy, enrolment and out-of-school population; Table 15, population and employment by age bracket. Islamabad. https://www.pbs.gov.pk/census

[2]  Pakistan Bureau of Statistics. Socio-Economic Indicators of Sindh at District Level. Islamabad. https://www.pbs.gov.pk/wp-content/uploads/2020/07/Socio-Economic-Indicators-of-Sindh-at-District-Level.pdf

[3]  Pakistan Bureau of Statistics. Population and growth rate by district, Census 1951 to 2023 (Table 1.02).

[4]  Government of Sindh. Sindh Right of Children to Free and Compulsory Education Act, 2013. https://www.sindheducation.gov.pk/Contents/Menu/Sindh_Bill_RTFE.pdf

[5]  Government of Sindh, Education and Literacy Department, Reform Support Unit. Sindh Education Profile 2013-14 (SEMIS). Karachi.

[6]  Government of Sindh, School Education and Literacy Department. Profiling for Government Schools 2019. Karachi.

[7]  Government of Sindh, School Education and Literacy Department / Reform Support Unit. Education Indicator: Annual School Census 2024-25. Karachi, August 2025.

[8]  Government of Sindh, School Education and Literacy Department. Girls Stipend Program — Overcoming a barrier to female education. https://www.sindheducation.gov.pk/pages.jsp?page=GirlsStipendProgram-Overcomingabarriertofemaleeducation

[9]  Government of Sindh, School Education and Literacy Department. Notification: girls' stipend eligibility criteria for the academic year 2024-25. https://www.sindheducation.gov.pk/Contents/Notifications/116235%20Notification%20for%20notify%20the%20girls%20stiped%20eligibility%20criteria%20for%20the%20academic%20year%202024-25.PDF

[10]  Worlddata.info, based on International Monetary Fund and World Bank series. Inflation rates in Pakistan. https://www.worlddata.info/asia/pakistan/inflation-rates.php

[11]  Government of Sindh, School Education and Literacy Department. District- and school-wise School Specific Budget, FY2025-26.

[12]  Hasan, S. 'Sindh Budget 2026-27: School-specific budget up despite smaller outlay for education sector.' Dawn, 18 June 2026. https://www.dawn.com/news/2008652

[13]  Naz, L. 'Sindh has increased its education budget by 25pc. But is it enough for half of its children who are out of school?' Dawn, 18 July 2025. https://www.dawn.com/news/1923019

[14]  Government of Pakistan, Finance Division. Pakistan Economic Survey 2024-25, Chapter 10: Education. https://www.finance.gov.pk/survey/chapter_25/Highlights.pdf

[15]  ASER Pakistan. Annual Status of Education Report 2023: Report Card, Rural Sindh. Idara-e-Taleem-o-Aagahi. https://aserpakistan.org/document/aser/2023/reports/Report-Card-Rural-Sindh.pdf

[16]  Humanitarian Data Exchange / OCHA. Pakistan — Subnational Administrative Boundaries (admin level 2 centroids). https://data.humdata.org/

[17]  Government of Sindh, School Education and Literacy Department. Sindh School Monitoring System (SSMS): objectives and indicators. https://www.sindheducation.gov.pk/pages.jsp?page=sindhschoolmonitoringsystem (accessed August 2026). See also UNICEF Pakistan, 'Mobile app helps reduce school dropout in Pakistan's Sindh province', supported by the European Union.

[18]  Government of Sindh. Student Attendance Monitoring and Redress System (SAMRS). https://samr.edu.pk/ (accessed August 2026).

[19]  'Clocking in gets a new face in Sindh schools', The Express Tribune, August 2026; and 'Sindh approves facial recognition attendance system for schools', ProPakistani, 4 August 2026. FRAMES pilot approved at a School Education Department meeting chaired by the Chief Minister; rollout in one district per division from 17 August 2026.

[20]  Asian Development Bank. Strengthening School Teaching in Pakistan. Manila: ADB, December 2023. Reported in Ahmed, A. 'Biometric monitoring reduces teacher absenteeism: ADB', Dawn, 28 December 2023. https://www.dawn.com/news/1801228

[21]  ASER Pakistan. Annual Status of Education Report 2019: Sindh. Idara-e-Taleem-o-Aagahi, 2020. Teacher attendance recorded at 89 % in surveyed government schools in Sindh on the day of visit.

A note on method

Out-of-school children, children who never attended, and dropouts are counted for ages 5-16 by Census 2023 and are used exactly as published. The denominator throughout is the census count of children aged 5-16 in each district (16,891,397 for Sindh), so every rate in this article is a true rate rather than a projection.

School counts, teacher counts, facility coverage and class-wise enrolment come from the Annual School Census 2024-25 and cover government schools only. Private-school enrolment appears only where the census reports it, at the primary stage. Because Karachi's schooling is overwhelmingly private, government-school indicators understate provision there, and that is stated wherever it affects an argument.

Correlations are Pearson coefficients across Sindh's 30 districts. They describe association, not causation, and the article says so wherever a causal reading would be tempting. Where a relationship depends heavily on a subset of districts — as the school-supply result depends on Karachi — the result is reported both with and without that subset.