Use of official statistics on ‘leading cause of homelessness’Make Text Bigger
Details of a complaint made to Statistics.Gov.Uk by Property118 member Owen O’Neil
I would like to make a complaint about simplification of the context of statistics to the point of misrepresentation that would mislead the public which falls under this duty of your office :-
(III) official statistics in a document or statement are presented in such a way that, in the Authority’s opinion, they are liable to mislead the public or undermine the integrity of official statistics.
The statistic that I feel is being misrepresented to the point of being misleading is ‘the main cause of homelessness’.
This has now appeared in two recent official documents, with varying quality of the presentation of the overall context of the statistic.
The most egregious is :-
Page 5 “The leading cause of homelessness is now attributed to the ending of an Assured Shorthold Tenancy (AST)”
Then following behind we have :-
“The ending of private sector tenancies has overtaken all other causes to become the biggest single driver of statutory homelessness in England. The proportion of households accepted as homeless by local authorities due to the end of an assured shorthold tenancy increased from 11% during 2009-10 to 32% during 2016-17. The proportion in London increased during the same period from 10% to 39%. Across England, the ending of private sector tenancies accounts for 74% of the growth in households who qualify for temporary accommodation since 2009-10. Before this increase, homelessness was driven by other causes. These included more personal factors, such as relationship breakdown and parents no longer being willing or able to house children in their own homes. The end of an assured shorthold tenancy is the defining characteristic of the increase in homelessness that has occurred since 2010 (paragraphs 1.14 to 1.16).”
(there is no mention that ‘statutory homeless’ is a narrow and selective subset of total homelessness)
A number of non-government organisations have made similar selective representations or contextual omissions when discussing the same subject.
The different authors vary in the quality of their description of the background, but none seem to provide adequate information as to the background to understand that this measure is a particular and restricted slice of homeless people that have been through a number of filtering processes. Given the political lobbying that a number of these organisations carry out, it would be reasonable to assume that their viewpoint and representation of the statistic is heavily coloured by their politics.
I have also been to a public meeting held by the Labour party where this throwaway ‘fact’ was stated without any kind of qualification.
This is a matter of public interest since it is being used as an argument for requesting a change in the law with respect to termination of a tenancy via use of a ‘section 21’ notice.
I fear if the context is not stated alongside the statistic then as the statement becomes repeated more often it will become an accepted and unchallenged ‘fact’.
To explain the context behind the statistic :-
Local authorities collect data on applications made to them….
Table 770: Decisions taken by local authorities under the 1996 Housing Act on applications from eligible households
So for starters just to make an application you have to be ‘eligible‘ – I don’t know what this criteria is, but at least some applicants will fall out before reaching this point…
|Total decisions 2017||112K||100%|
|Unintentionally homeless & priority need (acceptances)||58K||52%|
|Intentionally homeless & priority need||9K||8%|
|Homeless but not priority need||19k||17%|
Only if you make it through that stage and are accepted do they collect data on the cause.
Table 774: Reason for loss of last settled home:
Households accepted by local authorities as owed a main homelessness duty by reason for loss of last settled home
|Total acceptances 2017||58K||100%|
|Relatives/friends no longer able or willing||15K||27%|
|End of AST||16K||28%|
|Loss of other rented or tied||3K||6%|
It is this single table that all the statements are based on. We don’t know the quality / error rates for this data. For instance we do not know if there are secure tenancies (granted by RSL) being mis-reported as AST, also we do not know if the ‘rent arrears’ figure is accurate or if ‘end of AST’ also incorporates a significant number where rent arrears are the root cause. Given that failing to pay rent may result in being classed as intentionally homeless and hence excluded from acceptance it would be reasonable to expect that a former tenant would not be forthcoming to volunteer if rent arrears were the underlying cause. Furthermore ‘other’ comprises a significant amount of the cases, which have not been analysed to determine if this is bad data which should be assigned to another category.
Also note the difference between this and the next category (Relatives/friends no longer able or willing) is only 1% which is reasonable to suppose is within the margin of error of the data.
The primary categories of ‘Relatives/friends no longer able or willing’ & ‘Relationship breakdown’ are broken down into further sub categories within the published data – some authors further exaggerate the differences by comparing the amount of these sub-categories with the primary category of ‘end of AST’. Sadly page 7 of this parliament briefing paper falls into this trap.
Whilst it does fall into this trap the above parliament briefing paper is otherwise reasonably well balanced in placing the statistic in context, which all other authors fail to do.
“Local authorities in England have a duty to secure accommodation for unintentionally homeless households who fall into a ‘priority need’ category. There is no duty to secure accommodation for all homeless people. For example, there is no statutory duty to secure housing for homeless single people and couples without children who are not deemed to be vulnerable for some reason.”
“Organisations such as Shelter and Crisis argue that the official statistics do not give a full picture of homelessness in England. The figures exclude those who are homeless but who do not approach a local authority for assistance and those who do not meet the statutory criteria. . For example, of the 29,340 applications for assistance that English local authorities received between July and September 2017, only 52% were accepted as homeless and owed a main homelessness duty; the remainder were deemed either to be homeless but not in priority need, intentionally homeless or not homeless “
Turning to repossession claim statistics :-
Repossession claims started 2017
146K (mortgage 18K, social 78K, private+accelerated 50K)
We don’t know how many of these are resolved and the occupier remains in the property,
or how many never get into the process to start with because the occupier leaves when served with notice.
There are figures on how many make it further through the process, but we don’t know the reasons for claims falling out of the process, be that the matter has been resolved or the occupier has left.
Repossession orders granted 2017
109K (mortgage 12K, social 58K, private+accelerated 39K)
It would be reasonable to expect that a proportion of those will resolve their housing situation without contacting the local authority
However it is clear that social landlords are granted 50% more possession orders than private landlords (and the private sector is a significantly larger sector) – so to suggest that none of these people are made homeless as a result of eviction by a social landlord is implausible in the extreme. It is possible that (pro rata for size of housing sector) – evictions by social landlords cause more homelessness than evictions by private landlords – but we don’t have data on this, further analysis of the reasons for an applicant not being accepted might reveal this.
It would seem a more accurate statement to say that “of people made homeless, and who apply to a local authority, and are eligible, and are accepted as homeless (about half of applicants) – then amongst those people the ‘ending of an AST’ and ‘Relatives/friends no longer able or willing’ are within margin of error currently equal leading causes”
I hope you will review the use of these statistics and issue guidance on their correct interpretation to authors so that the statistics do not become devalued.
Please Log-In OR Become a member to reply to comments or subscribe to new comment notifications.