Nigel Hawkes: The real cost of a compromised census will be inaccurate data

Behind the Numbers

With less than a year to go to the 2011 Census, is Francis Maude about to change the rules? In opposition he was strongly critical of the Census, describing it as burdensome, intrusive and poor value for money.

Now he's in charge, as Minister for the Cabinet Office. He's also central to the programme of Government cuts that the coalition has promised. So demographers and statisticians are worried that he may make good on his promise to "scale back" the Census even at this late stage.

Censuses date back to 1801 and one has been held every decade since, with the exception of 1941. It's the only time that everybody in the country is counted, and it's a huge exercise. The 2011 Census will cost a whopping £482m, more than double the 2001 version (£207m) and the questionnaire dropping through everybody's door will be 32 pages long. Mr Maude is not alone in thinking that persuading everybody to complete it will be a tough task.

"How can the cost of half a billion pounds be justified at this time of fiscal crisis?" Mr Maude asked in the House of Commons in January. "In 2001, 10 per cent of the data was not even counted; it was imputed. Is this not a thoroughly wasteful and inaccurate exercise?"

He went on: "Should not a responsible Government be scaling the census back? Is not the answer a less intrusive, much cheaper census that offends the public less, increases compliance and therefore yields much more accurate information?" He made most of the same points in a pre-election meeting organised by the Royal Statistical Society (RSS) in February.

So, if he doesn't make any changes when he's now in a position to, he's going to have to eat a lot of words. He's already had a warning from the Demographics User Group, a private-sector body that represents heavy-hitters such as John Lewis, Marks & Spencer, Sainsbury's, Nationwide and Alliance Boots, saying that they regard the Census as of "fundamental importance" when making decisions about opening new stores, the products to be stocked, and the customers to be targeted.

A late decision to cut the number of questions would disrupt existing plans and contracts, save little money – or even add to the cost – and result in reduced and inconsistent information between England and Wales, Scotland and Northern Ireland, says Keith Dugmore, director of the group. The RSS is also understood to have added its voice, warning that any lack of enthusiasm at this stage is likely to damage the response rate.

It is not clear that much could be saved, anyway. The contracts were let long ago, including £150m to Lockheed Martin UK, for printing, data capture and processing, and £25m to Capita for training 35,000 temporary workers. The questionnaires are already printed, even though Census day is not until 27 March 2011 – a move which Mr Maude included in his list of Labour's "scorched earth" policies.

There are some serious risks that the Census will miss its targets. People are less and less willing to complete forms and an incomplete count means that areas such as London, which was undercounted in 2001, get less than their fair share of government funding. In future, probably before the 2021 Census is due, alternative methods of counting the population may be in place.

In that case, 2011 will have been the last-ever national Census. The risk is that Government indifference will implicitly give people the nod that they needn't bother with it, and then the money really will have been wasted.

Nigel Hawkes is Director of Straight Statistics; www.straightstatistics.org