Making Census?

For a researcher at Shelter, 28th June was something akin to the World Cup. New Census 2011 data showing housing tenure by family composition was released. In fact, it was more hotly anticipated, as we have had a ten year wait, and more cherished as it may be the last ever Census – more on that later.

First, what did it say? The number of families with dependent children renting privately has doubled from 600,000 in the 2001 Census to 1.2m. One in five families now rents privately, up from one in ten. This means that a third of privately rented homes now contain dependent children, up from a quarter in 2001. Shifts as pronounced as this, even over a decade, are extremely rare.

These trends are not new, but the Census proves them beyond any doubt and, crucially, delivers them to local authority, ward and even postcode level. We have been making the case for a private rented sector fit for families for a long time, but the Census numbers proved just how pressing this is becoming, with some areas seeing a 200% plus rise in the number of families renting over the decade.

That’s what makes talk of canning the Census particularly worrying.

So, is it, after 170 years, ‘over’, for us and the Census?

Mark Easton, from the BBC hosted a Social Research Association seminar on this very point a few days before the latest release. He showed us quotes from MPs in the 18th Century, to the modern day Daily Mail, all showing the Census to be roughly as popular as the plague. The anecdotes were amusing, but the point serious – the Census may be incredibly useful to social researchers, marketers and anyone needing to target certain groups in small local areas, but the British public have always been deeply distrustful of attempts to count them up and put them in boxes.

So, what is happening to the Census?

ONS assured us that all options remain on the table, including a ‘traditional’ 2021 Census. There will be a consultation on various alternatives. But, it felt to me as though there was pressure being exerted on ONS to move to something new.

So, what are the alternatives?

Government ‘Big data’ could potentially deliver the 95% coverage of the population that the Census currently achieves. And this would be annually. Sources would likely be the databases of the NHS, HMRC, DWP, and DVLA. Not many people are on none of these databases, and those that aren’t are probably roughly the same people who don’t fill in their Census forms.

But what about accuracy?

An audience member pointed out that his NHS card said he lived at his parent’s home, and his Driving License still had his previous rental address, and we all know this is not uncommon. The answer will lie in fully interlocking these databases to iron out most of the inaccuracies and achieve 95% accurate results for 95% of the population. But it is not known yet whether this is going to be possible by 2021. One thing that was made clear is that contrary to some tabloid reports, Tesco Clubcard data will not be the replacement for the Census!

But, the Census gives us far more than just local population counts, it tells us about how people are living in great local detail. However ’big’ the databases above are, they won’t tell us much about people’s housing situations. To fill this gap, large scale social surveys may be run every few years. To deliver reliable results on things like renting families by local authority, though, sample sizes will have to be close to 10 per cent of the population.

However you cut it, the Census would be hard to replace. For Shelter, truly big (and accurate) data on how, as well as where people live is vital to both service delivery and to tell the changing story of housing in England.