Call it is one of the unknown Indian ironies. Over many years, the Indian state, through its public irrigation agencies, has systematically taken over the management of surface water systems. It has taken over the job of building irrigation systems - dams, reservoirs and canals - then maintaining these and supplying water. This has meant that over the years it has taken over water resources from the hands of village communities. The irony is that even as the state has vested this power in itself, people have taken water under their control. Groundwater - a resource under the land of individuals and under their control - irrigates the bulk of lands in the country. State control is imagined. Indeed it is a delusion.

Currently, as much as three-fourths of the irrigated area is irrigated by groundwater, not surface systems. The irrigation infrastructure has been created by individual farmers - both rich and poor - using funds available from moneylenders or the meagre institutional finance provided by state credit agencies. It can be well argued that this lack of institutional support for infrastructure and the dependence on private finance is one key causes of farmer indebtedness and poverty in large parts of the country. The recently concluded 3rd minor irrigation census estimates that there are roughly 19 million wells - dugwells, shallow and deep tubewells - in the country. In other words, these decision-makers constitute the irrigation entrepreneurs of the country.

But to understand the governance of water we must understand why the economics of water is about the limitations of current technology. This is important because it has its reflections in other spheres as well - power generation, electricity supply etc. It is well accepted that investment in water infrastructure is confounded by both increased costs of irrigation infrastructure and declining investment from the state. The mid-term appraisal of the 10th Plan has estimated that the capital costs of extending irrigation can range from Rs 40,000 per hectare (ha) to Rs 2,50,000 ha, where storage facilities are required - which is over double from the previous decade. This is not even accounting for real costs of rehabilitating people who will be submerged or the biodiversity and forest that will be lost.

As the cost of infrastructure has increased, it has not been possible to recover the cost of capital investment and operation and maintenance of the system from individual farmers. In most states, agencies recover less than 30 per cent of the costs of maintenance of systems. This has led to their deterioration, with the bulk of surface water-fed canals in need of desperate repair. It is no surprise then, that the gap between the capacity of the irrigation infrastructure and its utilisation is assessed to be 14 million ha - as much as 20 per cent of the irrigation acreage of the country. In addition, large surface irrigation systems, which require transporting water over long distances, have substantial losses and inefficiencies. It is estimated that surface irrigation systems operate with as little as 35-40 per cent efficiency, while groundwater irrigation systems work at 65-70 per cent.

Because of this capital and resource intensity, it is not possible to be inclusive - reach everyone. Currently, over 45 per cent of foodgrain is grown on rainfed land, which sustains the poorest. In other words, investment in surface irrigation systems has created islands of prosperity but has done little to improve local food security.

This is where groundwater kicks in. It works because it is in the hands of people, who can use it when they need to irrigate. If there is no electricity to run their tubewell, they can buy diesel or rent a generator set to pump water. It is well accepted that groundwater use is more efficient - farmers use it more prudently. It is more cost-effective because it is more localised and transport costs are lower. In many ways, groundwater provides the most distributed and decentralised option for development. But only if managed wisely.

The intense use of this resource has meant that groundwater levels across the country are falling sharply. Technology is allowing for deeper and deeper penetration and extraction. The electricity subsidy - providing cheap energy for pumping - worsens the situation, with estimations that farmers end up using almost double the water for each unit of crop when they have access to cheap or free power as compared to pump-sets using paid diesel.

We can try and regulate this use with legislation. But clearly regulating the use of 19 million users will be difficult, if not impossible. What we have to recognise is that groundwater is a replenishable asset and what is needed is to recharge wells, so that annual extraction is limited to what is sustainable. In other words, we work groundwater like a bank. Live off the interest - what is recharged - and not the capital.

This is where the irony doubles. Even as groundwater has overtaken surface water systems, other irrigation systems - tanks, ponds and all other community-based and decentralised water harvesting systems - have simultaneously declined. But the fact is that these systems played a critical role in the recharge of groundwater as they stored rainwater, which then recharged underground aquifers. These were the 'distributed' sponges without which 'distributed' water management would not be possible. Therefore, we are extracting more and more, recharging less and less.

The tragedy is that when we lost respect for traditional systems which were designed to ensure that rainwater was stored in millions of disaggregated, diverse structures, we lost our water future. Clearly, we must understand the politics of technology to understand the economics that will matter.

- Sunita Narain