Pakistan publishes its agriculture census, revealing the federation beneath our fields

The state’s desire to count is never entirely innocent. British India turned censuses, land records and irrigation maps into technologies of rule: enumeration made people and property legible, while the canal colonies linked land settlement, water allocation and revenue to a powerful bureaucracy. Pakistan inherited that administrative apparatus. Yet the same numbers that can help a state extract can also help citizens see what is changing and demand a response.
The Agricultural Census 2024 deserves to be read in that critical but constructive spirit. It is Pakistan’s seventh agricultural census and the first to combine agriculture, livestock and farm machinery in one digital exercise.

This 2024 census is the first to combine agriculture, livestock and farm machinery in one digital exercise

Fieldwork was carried out in two phases between September 2024 and February 2025. It was sample-based rather than a literal count of every farm. (This was a different approach from work done before 2010 and from that of methods used by many countries, and therefore, its results merit more scrutiny). Selected mouzas and urban blocks were surveyed, while exceptionally large land holdings were included with certainty. Tablets, GIS mapping and real-time monitoring were used to produce district-level estimates. The numbers are therefore not sacred, but they are the most systematic national picture available.

More farms, smaller farms, and arguably lesser productivity

The most alarming finding is the continued fragmentation of land. Pakistan had 8.26 million farms in 2010; it now has 11.1 million, an increase of 34 per cent. Farm area grew by only 12 per cent, so average farm size fell from 6.4 to 5.3 acres. More strikingly, the number of fragmented farms rose from 2.83 million to 4.98 million, while the average number of separate pieces within a fragmented holding increased from three to seven.
Fragmentation is not a neutral statistic. Scattered parcels raise the time and cost of ploughing, irrigating and guarding a holding; they shrink the returns to levelling land, laying watercourses or investing in a tubewell that would pay off on a single consolidated field; they multiply boundary disputes between neighbours; and they push machinery, which needs scale to be worth renting or owning, further out of reach for smallholders. A farm split into seven pieces is not simply smaller. It is harder to farm well.
Inheritance is central to this story. After each generation, land titles are divided among heirs and the operational holding is often divided with them. Women are still frequently denied their lawful shares, so Pakistan manages to combine fragmentation with gender injustice. The answer cannot be to weaken inheritance rights. It is to separate ownership from operation.

Families should be able to retain distinct legal titles and income shares while cultivating adjoining parcels as one unit. Producer cooperatives, family farm companies, machinery pools, digital lease markets and voluntary land exchanges can create scale without dispossession. Government can provide cadastral maps, low-cost lease registration, model contracts and quick dispute resolution. Women heirs must be recorded as members, paid directly and given enforceable exit rights. Pakistan does not necessarily need fewer owners; it needs institutions that allow small owners to farm together.

From canal command to groundwater autonomy

The second transformation is taking place below the soil. Total irrigated area rose from about 34.1 million acres in 2010 to 45.9 million in 2024. Canal-only irrigation did increase, from 12.3 to 14.5 million acres, but it lost relative importance: its share of irrigated land fell from roughly 36 to 32 per cent. Land using both canals and tubewells slipped from 13.9 to 13.5 million acres. Tubewell-only irrigation, meanwhile, more than doubled from 6.1 to 14.1 million acres, raising its share from about 18 to 31 per cent.

This is not a story of canals disappearing. It is a shift in the centre of gravity from publicly coordinated surface water to privately controlled groundwater. The number of tubewells and lift pumps rose from 0.93 million in 2004 to 1.83 million in 2024.

It is a shift in the centre of gravity from publicly coordinated surface water to privately controlled groundwater.

The more striking shift is in what powers them. Solar barely registered as a power source for tubewells in 2004. Today it runs an estimated 960,000 of the country’s 1.83 million tubewells and lift pumps—roughly half the entire stock, and more than diesel and electricity combined. Diesel-powered units have fallen as fuel costs rose; grid electricity remains a minority option, constrained by an unreliable and expensive supply. In the space of two census rounds, solar has gone from a rounding error to the single largest power source irrigating Pakistan’s farms. This is one of the most consequential technology shifts in the recent history of Pakistani agriculture, and it has had almost no sustained political or policy conversation attached to it—no coordinated financing programme, no groundwater-metering requirement tied to the subsidy-free upfront cost that made it attractive, no basin-level plan for what happens when pumping is limited only by sunlight rather than by fuel budgets.

But private freedom can create a public crisis. A farmer who pumps groundwater receives the immediate benefit, while the falling water table, salinity and declining aquifer quality are shared by neighbours and future generations. Solar makes each additional hour of pumping appear almost costless. The transition therefore demands much more research: Where are water tables falling fastest? How much pumping is replenished? Who can afford deeper wells? How are groundwater and river flows connected?

By corollary, we need to study how this will affect cities. If agriculture becomes less dependent on canals, more surface water might theoretically become available for urban, industrial or environmental use. But that outcome is not automatic: Pakistan needs basin-level accounting before celebrating either liberation or surplus. The new autonomy of the farmer must be matched by collective rules for the aquifer.

The federation on the fruit cart

The census also offers a more hopeful way of seeing the federation. Pakistan’s unity is usually discussed through constitutions, dynasties and political threats. Yet its deeper roots may lie in grain sacks, milk tankers and fruit crates moving across provincial borders.

The tree counts show extraordinary specialisation. Punjab contains about 83 per cent of Pakistan’s mango trees, virtually all its kinnow trees and 77 per cent of its orange and malta trees. Balochistan has 94 per cent of the apple trees, 87 per cent of the pomegranate trees and about half the date trees. Khyber Pakhtunkhwa has approximately 95 per cent of the peach trees and 95 per cent of the walnut trees. Sindh accounts for about 38 per cent of the date trees and remains the second major home of mangoes—and it is the undisputed home of banana, growing 99 per cent of the national crop by cultivated area, the most complete specialisation of any province in any crop in this census.

The crop map is less evenly distributed than the fruit map. Punjab leads wheat, rice, cotton, sugarcane and maize alike, with Sindh a consistent second and Khyber Pakhtunkhwa a close rival only in maize. That concentration cuts both ways. It gives Punjab outsized weight in national food security, and it means a bad Punjab season — a flood, a heatwave, a canal-supply shortfall — is a national event rather than a provincial one in a way that fruit-tree specialisation, spread across four provinces, is not.

Inside that crop map sits a genuine alarm. Cotton area fell from 9.23 million acres in 2010 to 6.51 million in 2024, a drop of 29 per cent, even as the country’s total cropped area grew by 22 per cent over the same period. Cotton’s own share of the national cropped area collapsed from roughly 14 per cent to under 8 per cent. Pakistan’s textile sector, still the country’s largest export earner, was built on a domestic cotton base. A raw-material supply shrinking this fast while the mills it feeds keep expanding is not a sustainable arrangement; it points toward Pakistan becoming a structurally larger importer of raw cotton, spending scarce foreign exchange on a crop the country was once a major grower of.

There is a second pattern sitting uncomfortably next to the cotton collapse. Rice area grew by 15 per cent between 2010 and 2024 — a real absolute increase in the amount of land given over to one of the thirstiest crops Pakistan grows — even though rice’s share of the cropped area fell slightly because other crops expanded faster still. Sugarcane, another heavy water user concentrated in Punjab and Sindh, barely moved. Climate adaptation would normally argue for the opposite direction: shifting acreage toward crops that need less water per calorie or rupee of output, as rainfall becomes less reliable and summers grow hotter. Instead the census shows a country still expanding its most water-intensive staples in absolute terms, and financing the water for them increasingly from tubewells rather than canals — that is, from a groundwater account whose balance nobody is publishing.

Livestock is the faster-moving countryside

A final finding receives too little attention: livestock has expanded much faster than crop acreage. Between the livestock censuses of 2006 and 2024, cattle numbers increased by 89 per cent, buffalo by 75 per cent, sheep by 68 per cent and goats by 78 per cent. Taken together, these four major species rose by about 78 per cent. The number of in-milk cows increased by 140 per cent and in-milk buffalo by 111 per cent.

By comparison, between 2010 and 2024 cultivated area grew by 24 per cent, cropped area by 22 per cent and orchard area by 42 per cent. The periods and units are different, so this is not a direct productivity comparison. It nevertheless suggests that the livestock economy is becoming an increasingly important source of rural resilience, household assets, nutrition and cash income. Policy still treats livestock as an appendix to crops. Veterinary services, feed markets, breeding, disease surveillance, milk collection and cold chains should move much closer to the centre of agricultural planning.

A mirror, not a verdict

Every statistical exercise contains sampling error, reporting error, classification problems and institutional incentives. People who supply the raw data for a census (the respondents) forget or conceal; definitions change; governments highlight convenient findings. The census should therefore be questioned, cross-checked and opened to independent researchers.

Imperfect statistics are not made worse by scrutiny; they are made useful by it. The alternative is not perfect knowledge. It is policy by anecdote, lobbying and instinct. Pakistan’s new census shows a countryside that is fragmenting in ownership, gaining autonomy in water, integrating through food and shifting toward livestock. A census is only a mirror. The real test is whether the state merely counts these changes — or learns how to govern them.

Author’s data note

Figures are calculated from Pakistan Bureau of Statistics, Agricultural Census 2024 (National Report, released July 2026), with 2010 irrigation comparisons drawn from the corresponding PBS census tables. Provincial fruit figures refer to shares of counted trees; crop figures refer to shares of planted area. Livestock and crop comparisons cover different census intervals and should be read as directional rather than as a direct productivity comparison. All top-line figures were cross-checked against PBS’s primary “Agricultural Census 2024 at a Glance” summary table. Author’s Note on AI Use: In preparing this article, the author used Claude (Anthropic) as a research and drafting tool — for computing summary statistics from Pakistan Bureau of Statistics PDF reports, generating chart data tables, and cross-checking figures cited in the text against the PBS Agricultural Census 2024 Country Report and at places other reports. All analysis, interpretation, and conclusions are the author’s own. The author reviewed, verified, and takes full responsibility for all data, analysis, and conclusions presented.