Nigel Hawkes: Meters don't save water – they pump cash from poor to rich

For years, getting a water meter installed has been a one-way bet. For many households, it's the easiest way to save £100 a year, in some cases a lot more. All it takes is a phone call: the water company is obliged to install a meter, free of charge (except in Scotland or if it's impracticable, such as in blocks of flats with shared pipes) and then you only pay for the water you use.

The object of the policy, introduced in 1999, is to reduce consumption. People whose water is metered use less than those whose bill is calculated the old way, on the basis of the property's "rateable value". Rates haven't existed since 1990, so this really is a hangover from the past. If you have a meter installed and it doesn't save you money, you can opt to go back to the old system, so long as you do it within a year. It really is a one-way bet.

But is it good policy? Only if you believe that people with large houses and few occupants should be subsidised by those with big families living in smaller houses – put more crudely, the rich by the poor. Those with the greatest incentive for switching live in houses with high rateable values, but don't use a lot of water. As a rule of thumb, if your house has more bedrooms than people, you should switch.

The perverse results of the policy are due to a lack of data, or rather a lack of transparent data. Governments collect endless statistics about the way we live, but there are still corners of life where key pieces of information are available only to one partner in a bargain. This is the case with water metering, as Dr Simon Cowan of the University of Oxford explains in an article in the current edition of The Economic Journal.

The water regulator would like to design a system in which installing meters is cost-effective, so that the cost of installing the meter (around £200) is less than the value of the water saved. That means that water metering ought to be selective, rather than universal, and directed at households that use the most.

But the regulator does not know how much water an unmetered household uses, and lacks information on the size of the house, how many bedrooms it has, how many people live there, and what their income is. The people that live in the house don't know how much water they use, either, but they do know the rest.

So while social utility would best be served by metering families with children who use a lot of water and would respond by using less, what has actually happened is the opposite. Those who opt for metering are older or childless couples, people living alone, or those who know they don't use much water. They are the least likely to cut consumption, even if metered water prices rise. As things stand, they can use the same amount of water and save money, so the cost of installing the meter is borne by those less able to pay.

Dr Cowan's economic analysis is complex, and his conclusion careful. "This raises the possibility that current metering policies may be encouraging exactly the wrong households to choose to be metered," he says. "The free meters policy effectively lends households money at the water company's cost of capital, with the loan being paid back through higher tariffs." Except that the tariffs aren't higher: they're lower.

Water companies have so far spent around £500m installing meters. That's £500m transferred from the deserving to the less deserving, by my reckoning. And very little water saved.

Nigel Hawkes is Director of Straight Statistics (www.straightstatistics.org)