The Aquifer Is the Real Budget Constraint

Pritiraj BrahmaFounding Partner, Pritiraj & Partners · 26 March 2026

India extracts 247.22 billion cubic metres of groundwater a year — a quarter of the world’s total — and 730 assessment units take out more than comes in. The national trend is improving genuinely. Neither fact helps a farmer in one of those 730 blocks, because the state subsidises the electricity that pumps the water and procures the crops that require it.

ProvenancePublished 26 March 2026; revised 28 August 2026 to add two findings, two figures and a fuller counter-case. All groundwater figures are from the Central Ground Water Board’s National Compilation of Dynamic Ground Water Resources of India, 2025, with the 2017 comparison from the corresponding earlier assessment. The number of assessment units and elements of the assessment methodology have changed across editions, so category shares across years are not strictly like-for-like; this is stated in the essay rather than smoothed over. Programme figures for Jal Sanchay Jan Bhagidari, MGNREGA water works and NAQUIM mapping are as published by the responsible ministries.

India extracts 247.22 billion cubic metres of groundwater a year, roughly a quarter of the world’s total, and 730 assessment units now take out more than comes in. The national figures are improving genuinely and quickly. Neither fact helps a farmer in one of those 730 blocks, because India subsidises the electricity that pumps the water and then procures the crops that require it. Every part of that chain is separately defensible and jointly ruinous.

The National Compilation of Dynamic Ground Water Resources of India, 2025, assesses total annual groundwater recharge at 448.52 billion cubic metres, annual extractable resource at 407.75 BCM, and total annual extraction at 247.22 BCM — a stage of extraction of 60.63 per cent for the country as a whole.

The national average conceals the problem entirely, which is the first thing to say about it.

Six findings anchor this analysis:

1. Depletion is intense, local, and concentrated where it does most damage. Of 6,762 assessment units, 730 — 10.80 per cent — are categorised as over-exploited, meaning extraction exceeds annual replenishment. A further 2.97 per cent are critical and 11.21 per cent semi-critical. Punjab alone has 111 of its 153 blocks over-exploited, the second-highest count in the country after Rajasthan.
2. Almost everything depends on it, and almost all of it goes to one sector. Groundwater sustains about 64 per cent of irrigation, 85 per cent of rural drinking water and about half of urban water needs; some 87 per cent of what is extracted goes to agriculture. India is the world’s largest consumer of groundwater, accounting for roughly a quarter of global extraction. This is not one resource question among several — it is the substrate of Indian food security and rural drinking water simultaneously.
3. The national trend is genuinely improving, and the improvement is real. Recharge has risen from 432 BCM in 2017 to 448.52 BCM in 2025. The share of safe assessment units has risen from 62.6 to 73.14 per cent, and over-exploited units have fallen from 17.2 to 10.8 per cent across the same period. This is one of the few environmental indicators in India moving decisively in the right direction, and the essay says so before criticising anything.
4. Improvement at the aggregate does not help a block already past 100 per cent. A unit extracting more than it recharges is drawing down stock, not flow. Once the water table falls below the reach of shallow wells, the cost of extraction rises for everyone and falls hardest on the smallest farmers, who cannot afford to deepen a borewell. Depletion is therefore regressive as well as unsustainable.
5. Indian law makes groundwater a common pool with private extraction rights. Groundwater is tied to land ownership: the owner of the surface may extract beneath it. There is no effective mechanism by which one farmer’s over-extraction constrains another’s, which is the textbook condition for depletion of a shared resource. No amount of recharge construction alters that property rule.
6. The instruments now being used are the right ones. Jal Sanchay Jan Bhagidari targets one crore artificial recharge and storage structures; 65 per cent of MGNREGA funds in over-exploited and critical blocks have been earmarked for water-related works, and fifteen additional recharge activities — recharge shafts, injection wells, filtration ponds, artificial wetlands, spring collection chambers — have been permitted under MGNREGA for stressed districts. Converging an employment guarantee with an aquifer problem is intelligent policy design, and it addresses supply.

Where the water is actually running out

Recharge structures address supply. The harder half is demand, and demand is shaped by three policies that no state finds easy to touch.

Free or flat-rate agricultural power. When the marginal cost of pumping is zero, the rational farmer pumps until the crop is saturated. This is not misbehaviour; it is the response any economic actor would have. Metering agricultural power is the single most effective demand-side intervention available, and it is close to politically untouchable in most states.

Assured procurement of water-intensive crops. Where paddy and wheat have reliable procurement at a support price and pulses, millets and oilseeds do not, cropping choice follows procurement rather than hydrology. Punjab’s block data is the arithmetic result of that policy.

A property rule that cannot be enforced against. Because extraction rights follow land, a village cannot restrain the neighbour whose borewell is drawing down the shared water table, and a state that wanted to would have to legislate against a right attached to ownership.

The direction of travel, honestly reported

It is worth dwelling on this, because the standard environmental essay would not. The recharge programme is working. India has moved a hydrological indicator in the right direction across eight years, at scale, using an employment guarantee as the delivery vehicle. That is a serious governance achievement and it demonstrates that this problem is tractable rather than fated.

The counter-case, honestly stated

Three arguments cut against this analysis, and the last is against our own remedy.

First, groundwater is what made the Green Revolution work and what has kept Indian food security intact through drought years. It is farmer-controlled, requires no state irrigation bureaucracy, and delivers water on demand — advantages no canal system has matched. Restricting it without providing an alternative is a policy that reduces both output and rural incomes. That is right, and it defines the shape of a workable answer: substitution rather than restriction.

Second, part of the measured improvement may be methodological rather than hydrological. The number of assessment units and elements of the assessment method have changed across editions, so the category shares are not a strictly like-for-like series. The direction is consistent across intervening assessments, which is why we report it as real — but an essay that criticises the concealing power of a national average should not lean on a national average as its evidence of progress without saying that the series was rebuilt underneath it.

Third, and most seriously: feeder separation with metering and direct benefit transfer is our central recommendation, and its Indian evidence base is thinner than its logic. Several states have separated feeders; the step from measurement to pricing has stalled almost everywhere it has been attempted, and where cash-transfer pilots have run, participation has often been low and reversal politically swift. The mechanism is sound and the implementation record is poor, which means the honest recommendation is to publish the extraction data first and build the political case on it, rather than to present metering as a reform awaiting only courage.

What we would do

1. Publish the block-level balance annually, in plain terms. Every panchayat in an over-exploited block should receive one page each year: this is what recharged, this is what was pumped, this is the difference. Community management works where the community can see the number, and this is the precondition for every harder measure below rather than an afterthought.
2. Separate the feeder, then measure before pricing. Separating agricultural feeders allows supply to be scheduled and metered without touching domestic supply. Metering need not mean charging immediately — measurement first, pricing later, with the subsidy paid as direct benefit transfer rather than as free power. A farmer who receives the subsidy in cash and pays for what is pumped has an incentive to pump less and keep the difference. The record says this stalls at the pricing step; the answer is to bank the measurement gain regardless.
3. Shift procurement, not exhortation. Crop diversification campaigns fail against assured procurement. Procurement of pulses, millets and oilseeds at scale in over-exploited blocks would move cropping patterns in one season, because it changes the arithmetic rather than appealing to it. This is the single most powerful demand-side lever the Union holds and it requires no new law.
4. Manage at the aquifer, not the district. Aquifers do not follow administrative boundaries. NAQUIM has mapped the entire mappable area of the country, roughly 25 lakh square kilometres. That mapping should be the unit of management, with participatory water budgeting at village level against a published aquifer balance.
5. Put water in the industrial siting decision. Data centres, thermal plants and water-intensive manufacturing are being sited without a binding water test. In a stressed block, that decision commits a resource the block does not have, and it commits it for decades against a licence granted in months.

India’s fiscal deficit is debated every February. Its hydrological deficit — extraction beyond replenishment across 730 blocks — is not debated at all, and it is the constraint that will bind first for tens of millions of farming households. The recharge numbers show India can move this. What remains is the half of the problem that involves telling people what to pump and what to grow, and the block-level page that would let them see why.