Ministers were under fire today for failing to find a solution to the growing cost of social care.

With an ageing population, there are fears that the cost to the state of care home places for the elderly will grow to unsustainable levels.

But critics say it’s unfair that only the poorest pensioners get state help. Many more don’t qualify for financial support and we know that people are being forced to sell their houses to pay for the cost of accommodation in care homes privately.

But how many?

The government plans to give state loans to pensioners so they don’t have to sell their homes immediately – the money will be recouped from sales after death instead.

You would think ministers would want to know how many people are likely to need a loan like that before announcing the new policy, but we’ve seen wildly different claims about the situation in the media today.

According to various reports, anything from 20,000 to 40,000 people are forced to sell their properties every year to pay for care. What is the real extent of the problem?

The genesis of the “20,000 a year” claim appears to be this Daily Mail article from 2010.

The piece says that “at April 2010, some 47,000 have sold their home specifically to pay for care costs”.

“By using the fact that the average care home stay in England is 26 months, it has been possible to estimate that between April 2009 and March 2010, around 21,700 sold their home – 59 every day.

“This compares to 17,100 in 2001/2 and 18,500 in 2004/5.”

This, the Mail said, was all based on research from the House of Commons library and healthcare consultants Laing & Buisson.

But when we contacted both those organisations, they distanced themselves from the claim, saying they had never made any assessment of how many homes are sold each year to finance healthcare.

The Commons library did carry out some research for the Conservative MP Philip Davies, which appears to be the source of the Mail’s story, but Mr Davies’ office has thus far been unable to furnish FactCheck with the original numbers.

Assuming the figures are true, the Mail article seems to be saying that there were a total of 47,000 people living in care homes when a snapshot survey was done in April 2010 who had sold their homes to pay for their care.

If all of those people were in care for the average length of time (26 months), then you would get a complete turnaround of people in just over two years.

So about 21,700 new people who had sold their homes would enter the system each year. We think the Mail has divided 47,000 by 26 then multiplied it by 12 to get an annual figure.

What’s wrong with this?

First of all, the 26-month average, we presume, is for the whole population of people in care homes. That will include people who have been given means-tested state help and people who are able to finance their own stay.

It seems that “self-payers” – the group that includes people who have sold their houses, although not all of them will have done so – tend to stay in care for longer than average.

That could simply be because people with relatively more money tend to be healthier and live longer, but the author of the study, Dr Julien Forder, thinks it’s more likely to be because self-payers tend to go into homes earlier than people who rely on council-run services.

This report by the multi-university Personal Social Services Research Unit (PSSRU) found that “self-payers” in a cross-section of nursing and residential homes stayed for about ninth months longer than average. Their average length of stay was about 34 months.

So if you wanted to do a Mail-style calculation with these figures you would divide by 34 not 26, cutting that “20,000 a year” figure significantly.

More importantly, this kind of sum is probably just be too crude to be realistic.

Despite being fairly unreliable, the “20,000 a year” figure has cropped up in countless news reports on care homes.

Today, though, it suddenly doubled, with many news sources now saying “around 40,000” people are being forced to sell their homes each year.

More promisingly, the source for that number is the Department of Health. They cite as their source this study commissoned by the BBC in 2006 and written by health and social care consultant Melanie Henwood.

It contains a mini-FactCheck of the 40,000 number, concluding that it probably comes from research done 17 years ago “on the basis of some very broad assumptions about home ownership and entry to residential care”.

Ms Henwood notes that Lib Dem MP Paul Burstow estimated that 70,000 people were selling their homes every year. He got that number by obtaining figures from a small number of councils and extrapolating the percentage of people known to have sold their homes (16 per cent) nationally.

A government response to the Health Committee in 2005 suggested that about 30,000 people a year were selling their homes to fund care.

This all led Ms Henwood to conclude: “These various estimates suggest a range of figures on the likely numbers of properties that have to be sold each year to pay for residential care. Precise figures are unavailable because information is not collected centrally.

“However, it would seem reasonable to conclude that the number of homes sold each year to pay for care probably is somewhere in the region of 40,000.”

The verdict

We’re pretty dubious about the Mail’s “20,000 a year” story but we don’t know whether the real number is bigger or smaller – and neither does the government.

As far as the new estimate of 40,000 a year goes, we can’t in all honesty take things much further than Melanie Henwood.

She concluded that it wasn’t an unreasonable guess, but despite what you may have read today, it has no basis in hard fact.

 By Patrick Worrall