Errors, damn errors, and statistics: Ajilore v Hackney LBC

With the Supreme Court set to look at priority need this December, Ajilore v Hackney [2014] EWCA Civ 1273 may prove to be a brief footnote in the evolution of the bloated Pereira test. But, at least for the next four weeks, it tells us something about the construction of the ‘ordinary homeless person’ against which, post Johnson v Solihull [2013] EWCA Civ 752, applicants for homeless assistance are assessed.

Mr Ajilore had a history of drug addiction and crime and in 2010 he had served 18 months in prison for possession of Class A drugs. He suffered from depression and claimed to have made as previous suicide attempt. In prison he was said to have been on suicide watch His doctor reported that he had spoken of hopelessness and a desire to die. In his homeless application he stated that if he were made street homeless, he would be at risk of committing suicide, and at risk of relapsing into cocaine abuse.

A review of that decision found that the fact Mr Ajilore would be at risk of committing suicide due to depression, or of relapsing into drug use, did not differentiate him from other ordinary homeless people such as to make him a vulnerable person in priority need for housing. In so doing the review officer relied on statistics:

  • A report published in the British Medical Journal in 2005 which was said to confirm that homeless people have rate of self harm that was 7.2% higher.
  • A Joseph Rowntree Report on homelessness which had found that 38% of homeless people attempted suicide.

The review decision stated that “this shows that even if you did have suicidal thoughts that this would not necessarily be anything different to what an ordinary homeless person would suffer from.”

The county court accepted that the review officer had misunderstood or misrepresented the statistic on self-harm. In fact, the ‘report’, which was actually a letter to the journal from a doctor in a Leeds clinic, merely reported that the number taking overdoses in her clinic between 1999 and 2003 was “7.2%, [note the comma] higher than in the general population”. It was accepted that this misinterpretation had led to an inflated picture of the risk of suicide amongst the homeless [29]. Other statistics were said by Mr Ajilore to be misleading, for example the Joseph Rowntree statistic was taken from a survey of a self-selecting subset of the very most vulnerable homeless people, and in addition made no reference to the stage at which the individual had attempted suicide. The county court however, dismissed Mr Ajilore’s appeal against the review decision.

On appeal to the Court of Appeal Mr Ajilore submitted that (1) the review officer had erred in construing the comparator by misinterpreting the statistical evidence and inflating the risk of suicide amongst the homeless; (2) the reviewing officer had erred in holding that a risk of relapse into drug use was not capable of being a detriment for the purposes of the Pereira test, or alternatively, that the reviewing officer had failed to give adequate or sufficient reasons to explain how he had dealt with the question of drug relapse, and had not considered the impact of Mr Ajilore’s depression on his ability to avoid any relapse whilst homeless; (3) the reviewing officer had erred by comparing Mr Ajilore’s vulnerability with that of the hypothetical homeless person rather than with that of the hypothetical ordinary person, although it was accepted that Johnson was binding.

The appeal was dismissed.

(1) Although the fact that the officer had effectively misread the statistic and overstated the incidence of self-harm amongst the homeless, this did not undermine his conclusions that he was not satisfied that Mr Ajilore was genuinely seeking to harm himself, and that his risk of self-harm and suicide, if he were street homeless, would not be anything different from what he would expect to find in ordinary street homeless people.

The material before the officer confirmed that homeless people had higher self-harm incidence than the ordinary population. The statistics relied on by the officer were used to back up the conclusions he had drawn from his own knowledge of an ordinary street homeless person, as a skilled and experienced housing officer, and from his own assessment of the evidence relating to Mr Ajilore. The review decision was not based wholly or principally on the statistics. Accordingly, the officer’s errors in relation to, or arguable misinterpretation of, the statistics, had not permeated his reasoning so as to vitiate his whole decision (see paras 34-36, 39-42 of judgment).

(2) The officer’s statement that there would be a risk of relapse if J were made street homeless, but not to an extent that differentiated him from the ordinary street homeless person, was an implied recognition that, in some circumstances, the risk of relapse might be capable of constituting a detriment, but that, on the facts of J’s case, it was not detrimental or sufficiently detrimental to make J vulnerable. On the evidential material before him, the officer was clearly entitled to conclude that J’s risk of relapse did not make him any more vulnerable than the ordinary homeless person.

(3) The court was bound by Johnson v Solihull MBC [2013] H.L.R. 39, and it was not open to argue that the correct comparator had to be the ordinary person in the local authority’s district (para.55).

 

Comment

As the gatekeepers of social housing there can be no doubt that homelessness officers with dwindling budgets are under pressure to allow as few applications as possible. Nowhere is the commitment to protecting scarce resources felt more keenly than in cases where vulnerability is in issue.

The reliance on the heightened threshold of vulnerability endorsed by the Court of Appeal in Johnson v Solihull has proved one of the most insidious devices for rejecting otherwise plainly vulnerable applicants. On this analysis the ‘ordinary homeless person’ is a member of a subset of the general population, a subset that is, almost by definition, already vulnerable. By comparison, the applicant is run of the mill and, through the lens of Pereira, therefore not considered vulnerable at all. Such device delivers long term results for hard-pressed local authorities by creating a rising hurdle of vulnerability for future applicants: as the more extreme decisions force mentally ill, trauma-damaged and learning disabled applicants into street homelessness, so the average worsens.

In this approach, the ‘ordinary homeless person’ is fashioned by review officers as a product of their experience, and through the use of carefully selected statistics demonstrating the many problems experienced by those who find themselves street homeless. One might think that where the risk to the applicant is stated only as relative to this standard, an error in the benchmark would invalidate that risk assessment. However, that was not the conclusion supported by the Court of Appeal, which found that the statistics were used not to provide an empirical basis to the comparative exercise, but to “back up the conclusions that he had drawn from his own knowledge of an ordinary street homeless person, as a skilled and experienced housing officer” [39]. Thus, the statistical error didn’t necessarily matter as it was merely ancillary to the review officer’s assessment of norms among the street homeless. There is an obvious danger here that the ordinary homeless person is such a slippery concept that, no matter how desperate the applicant’s circumstances, it will always be just out of reach. If we were being cynical one might even think that was the intention.

Of course all of this takes the vulnerability assessment one step further from what was intended in the legislation. The purpose of the Act was to ensure that people who are at more risk of suffering harm when homeless are given accommodation. It is difficult to imagine that Hobhouse LJ, when he adopted the Pereira test, ever imagined that its use would result in a quasi statutory test whereby applicants who, on an ordinary meaning of the word are vulnerable, and are at risk of serious harm if made street homeless, are deemed not in priority need.

About S S

SS is a barrister practising in housing, discrimination and related public law.
Posted in Homeless, Housing law - All and tagged .

One Comment

  1. Pingback: Errors, damn errors, and statistics: Ajilore v Hackney LBC – NearlyLegal | Current Awareness

Leave a Reply (We can't offer advice on individual issues)

This site uses Akismet to reduce spam. Learn how your comment data is processed.