The Government’s rough sleeping statistics received a lot of attention when they were published in February this year.
This was because they reported a fall in rough sleeping, and because media and charities questioned their robustness.
This Insight looks at where the rough sleeping statistics seen in the media come from, the challenges of collecting them, and some of the quality issues that have been raised.
What are the statistics measuring?
The Government figures are intended to be an estimate of the number of people sleeping rough on a single night in autumn in England. They are based on data from local authorities and are for England only. The Government describes this as a ‘snapshot approach’ that isn’t intended to cover everyone with a history of sleeping rough, or other forms of homelessness such as ‘sofa surfing’.
The Government asks local authorities to capture data that describes one night between 1 October and 30 November.
Not all local authorities do this on the same night, though there is some co-ordination to avoid double-counting people. People are only counted if they are sleeping rough or about to bed down for the night outside – so the figures don’t include people in night shelters or hostels, or people who are without a place to stay but not sleeping in open spaces.
The count is timed to avoid the winter, when rough sleepers are more likely to be in temporary night shelters. Homelessness charity Glass Door has said the exact timing of the count can nevertheless affect the figures. It told Big Issue magazine:
“We know that many of the [London] borough-coordinated street counts took place after Glass Door and many other charities had opened their shelter doors to rough sleepers for the winter.”
What did the data for 2019 show?
Local authorities recorded 4,266 people sleeping rough in England for autumn 2019. This is a 9% decrease on the 4,677 recorded the year before. However, it’s a 141% increase on the number recorded in 2010, the first year in which this method was used.
1,136 people were recorded sleeping rough in London, representing 27% of the England total. The South East made up a further 21% of the total, meaning that almost half of all rough sleepers recorded were in London or the South East.
The Ministry of Housing, Communities and Local Government (MHCLG) has published a summary report alongside the data, describing regional variation and trends over time.
The Government’s official statistics aren’t the only figures available on rough sleeping. In London, the Combined Homelessness and Information Network (CHAIN) database is used by agencies working with rough sleepers to record contact they have with people affected.
What alternative figures are there?
The Greater London Authority publishes an annual report based on CHAIN data which details the number of rough sleepers ‘encountered by outreach workers’ in London across the year.
The 2018/19 annual report found that 8,855 people were seen sleeping rough in London during that year. This was an 18% increase on the previous year, despite a 7% fall recorded a year previously.
Of the people counted, 38% had been seen sleeping rough in previous years, while 62% were seen for the first time in 2018/19. Not all were regular rough sleepers: 60% were only seen rough sleeping in London once.
The CHAIN data has its own drawbacks. Dominic Williamson, an executive director at St Mungo’s, the charity that administers the database, told Inside Housing that the number of rough sleepers recorded is in part affected by the number of outreach workers available on the streets.
This means the increase reported in 2018/19 could be influenced by an increase in outreach workers available to work with and record people sleeping rough.
‘Counts’ versus ‘estimates’ in the official statistics
Local authorities can use different methods to produce the official statistics.
Some carry out a count; sending staff onto the streets to directly take a headcount of the number of people sleeping rough or bedded down.
Others produce an estimate by working in collaboration with local agencies that work with rough sleepers – such as the police, outreach workers, and addiction support agencies.
It’s not clear that one of these approaches is always better than the other. Homeless Link, the charity that supports and verifies local authorities’ figures, says:
“A local authority may choose to do a physical count if there is a high number of people sleeping rough in the area and sleep sites are both visible and easily accessible, such as urban areas.
Likewise, if there has been a significant change in the number of rough sleepers and there is a lack of evidence about who individuals are, local authorities may undertake a count. This is especially useful if there is a disagreement between partner agencies in the area about who is sleeping rough and where.
…Areas may choose to do an estimate if the local context of rough sleeping makes physically counting impractical (e.g. in large rural areas or areas with rough sleepers in areas that are inaccessible or unsafe to go into).”
Local authorities that carry out a count risk missing people who are not yet bedded down, or who are in inaccessible areas – so if partner agencies are able to help, an estimate may give a more comprehensive figure. But estimates carried out in areas where partner agencies don’t have good data on rough sleeping can also end up leaving people out.
Some local authorities combine the methods by taking a count and combining the figure with intelligence from local agencies. This is still officially recorded as an estimate.
Does switching between methods matter?
Concerns have been raised about local authorities switching between estimates and counts in different years. In 2017, 16% of local authorities carried out a count. This rose to 22% in 2018 and 25% in 2019.
A change in method makes it harder to tell how much rough sleeping has changed in an area. This makes it harder to evaluate how much the Government’s Rough Sleeping Initiative (RSI) funding has helped local authorities reduce the problem.
In a letter to the Government in April 2019, the UK Statistics Authority, the independent body that regulates the use of official statistics, expressed concern about this shift.
It said the statistics alone “should not be used to draw firm conclusions about recent trends in rough sleeping and cannot yet support public claims about the success of the Rough Sleeping Initiative.”
MHCLG has since published an evaluation of the Rough Sleeping Initiative in 2018 that attempts to account for the shift in methods used by some local authorities receiving funding between 2017 and 2018.
It found that rough sleeping did appear to decrease overall in areas that received RSI funding, compared with trends in rough sleeping in a control group that didn’t receive funding. However, the control group had lower levels of rough sleeping overall than the RSI group, meaning that it wasn’t completely comparable.
Will rough sleeping statistics improve?
MHCLG, the Office for National Statistics (ONS), and other government departments are working to improve statistics on homelessness and rough sleeping in general. This includes plans for a new types of research and evaluations. For example, the ONS has recently started to publish statistics on deaths of homeless people in England and Wales.
The Government’s current method for measuring rough sleeping is likely to be used for the foreseeable future, though. The homelessness charity Crisis recently called for a CHAIN-style system to be used across England, but MHCLG has previously said that it doesn’t consider this to be a practical option. However, it also says it is encouraging local authorities to improve the data that they collect year-round.
Further reading
Statutory Homelessness in England, House of Commons Library.
Rough sleepers: access to services and support (England), House of Commons Library.
About the author: Cassie Barton is a statistician in the House of Commons Library, specialising in demographic statistics.