Aquifer depletion

Why does an aquifer keep falling when every farmer drawing on it wants it to last?

Ask the farmers over a declining aquifer what they want and you get a remarkably consistent answer. They want the water to last. They can read the meter on their own well; they know the water table is dropping; many can name the year their fathers stopped hitting water at forty feet. Nobody is confused, and nobody is indifferent.

And the water table keeps falling. That is the part that deserves attention. We are used to explaining bad outcomes by ignorance or by bad actors, and here we have neither. So the explanation has to lie somewhere other than in what people know or what they want — which means the situation itself must be arranged so that informed, well-meaning people produce a result none of them chose.

Start with the physical fact that makes the arrangement possible. An aquifer is one connected body of water beneath many separately owned fields. Pump from your well and the cone of depression spreads outward under your neighbours' land; the water you remove is not removed only from beneath you. Two features follow, and both matter. The resource is subtractable — your litre is unavailable to anyone else — and it is hard to exclude anyone from, because a well on their own land is entirely within their rights. That pair of features is what defines a common-pool resource.

Now reason from a single farmer's position, and reason honestly rather than assuming greed. Suppose he considers pumping less this season to conserve. What does he actually get for it? The saved water does not stay in a personal account. It sits in the shared body, where his neighbours' wells can reach it — and next season it may well be theirs. Meanwhile the cost of restraint is entirely his: a smaller crop, this year, on his own balance sheet.

So examine the payoffs he faces. If the others restrain themselves and he does not, he does well. If the others pump hard and he restrains himself, he takes the loss and the aquifer falls anyway. Whichever the others do, pumping is his better move. Notice we did not need to make him selfish or short-sighted to get here. We only needed his savings to be capturable by someone else.

Push one step further, because groundwater adds something a fishery does not have so starkly. Deep aquifer water in arid regions recharges over centuries or millennia; the southern High Plains receives on the order of millimetres a year while irrigation withdraws vastly more. Water left in the ground is therefore not a renewable stock quietly regrowing — it is an asset that will be extracted by someone eventually. Leaving it is only worthwhile if you will be the one to get it later, and the more neighbours you have, the less likely that is. The effect is to raise everyone's effective impatience: the rational move is to pump it now, and the result is sometimes called the race to the pump.

We can now say precisely what is going wrong, and it is not a failure of understanding. The cost of one farmer's pumping falls mostly on others — a lower table, a longer lift, a higher energy bill, a dry shallow well next door. Each decision is individually sensible and collectively ruinous, because the decision-maker bears a fraction of what the decision costs. This is a structural property of the situation, so telling people about it changes nothing; they knew.

That diagnosis carries its own prescription. If the problem is that private and shared costs differ, then information campaigns, exhortation, and better technology will not fix it. Efficient irrigation can even make it worse — if drip systems let each acre use less, the freed water tends to go into irrigating more acres, so basin-wide withdrawal rises. Only something that changes what each farmer faces will do: a limit, a price, a right that can be held and traded, or a rule the community writes and enforces on itself.

The obvious conclusion is that depletion is inevitable, and the evidence says it is not. Elinor Ostrom's work, which won the 2009 economics Nobel, documented commons that have been governed successfully for centuries, and drew out what they had in common: clear boundaries on who may draw, rules matched to local conditions, users participating in making those rules, monitoring that users can verify, graduated sanctions rather than a single harsh penalty, and cheap dispute resolution. What these deliver is not virtue but credible mutual assurance — each user restrains because they can see the others doing so.

The groundwater cases fit. In Sheridan County, Kansas, irrigators themselves petitioned to create a local enhanced management area in 2013, agreeing to a roughly twenty percent cut in pumping enforced by metering; decline slowed sharply and farm incomes held up, because the water saved went to higher-value use rather than being lost to a neighbour. California's 2014 groundwater legislation attempts the same thing at a larger scale, requiring basins to reach sustainable management within decades. Both work by the same mechanism this reasoning predicted: they change the payoff, not the beliefs.

Two honest caveats. Enforcement is the hard part — unmetered wells make monitoring impossible, and where electricity for pumping is heavily subsidised, the price signal points the wrong way. And "sustainable" is a choice, not a physical constant: for an aquifer with negligible recharge, the real decision is how fast to spend a finite asset and what to leave the region afterwards.

• Why efficiency improvements often raise total water use, and what the rebound effect implies for irrigation subsidies.
• Land subsidence over pumped basins, and why the lost storage does not come back.
• How groundwater markets and tradable pumping rights perform where they have been tried.

Every term the collection defines is gathered in the glossary.