Relaaaaaaax, just have (blind) faith in your RAS

[Note: The research referred to in this post has been expanded upon since it was written and published in the Journal of Social Welfare and Family Law (Series, L. & Clements, L. (2013) 'Putting the Cart before the Horse: Resource Allocation Systems and Community Care ', Journal of Social Welfare and Family Law, 35(2) 207-226).  A pre-print version of this paper can be downloaded from here]

Last year I wrote a piece about transparency and personal budgets, which has had consistently high traffic so I assume it's a piece of general wider interest.  Since I conducted that initial research last summer, I embarked upon a more careful piece of research which sought to unpick different Resource Allocation Systems (RAS's) of several different local authorities.  I'm hoping, at some point, that my findings on RAS's will be published 'properly' and I don't want to give away too many spoilers until that paper is ready, but I did just want to comment on the ongoing transparency issues I've faced.  

In general, since my last piece of research, local authorities have been more willing to disclose their RAS algorithms to me.  I have a funny feeling that putting in my FOI request letter that it was for research into transparency helped somewhat.  In the event, only two refused disclosure - Croydon and Hackney - but their reasons for refusal pose wider democratic and accountability question about the RAS, which I want to consider in more detail.

Croydon: the contradictions at the heart of the RAS

Croydon's refusal was a pretty bog standard 'we're not going to tell you because we don't trust our service users not to fiddle their scores':

Upon request we provide customers with quite precise details of how our RAS tool has calculated their indicative allocation. This of course is always specific to their circumstances and subject to a full explanation of the purpose and role of an indicative allocation. However, there are a number of reasons why it would be inappropriate for us to share or publish our precise formulas in their totality. The main ones are:

1) The function of the RAS tool is to generate an indicative allocation to start support planning. Our RAS policy (which you have been sent) provides a far clearer and fuller explanation of our resource allocation system.

2) If people completing the form (or advocating) were aware of how the answers impacted on the indicative allocation, then they may be incentivised to provide inappropriate answers to try and get a higher indicative budget. This could have serious implications for our assessment and support planning processes; and potentially even threaten the sustainability of our adult social care budget.

3) The work of adult social care is in achieving personalised outcomes for individuals which are sustainable. The best way to promote this overriding objectiveis to focus on the community care assessment in the first instance. Over-emphasis on the RAS tool is very unhelpful from an operational and financial perspective (both generally and in individual cases).

4) The RAS tool is developed and maintained iteratively based on learning from practice. Changes to the RAS tool can be made at any time it is felt appropriate. It is the RAS Policy that represents Council Policy, not the tool itself. The aim of these changes is to improve the reliability of the indicative allocations so that they can better inform support planning, etc. It is important to remember here that the RAS tool does not “allocate” anything. The production of an appropriate, outcome focussed support plan is the mechanism / system by which resources are allocated.”

Croydon's refusal speaks volumes about the inherent paradox of Resource Allocation Systems.  In point 2 above, Croydon argue that knowing how the RAS produces indicative allocations might lead to people fiddling their results to they got a higher indicative budget, which could threaten the sustainability of their budget.  But this would suggest that the 'indicative amount' arrived at by the RAS is fixed and immutable, when we know that it is not.  If people do 'fiddle' their responses, proper input from care managers in the assessment process should pick this up anyway.  Croydon then go on to say that the RAS 'does not "allocate" anything' - they aren't an 'entitlement at all - which is precisely the reason why the argument made in point 2 cannot stand.  They also argue that 'over-emphasis on the RAS tool is very unhelpful from an operational and financial perspective' - I can hardly see why that is not a reason to disclose it.  

On the one hand, Croydon want to imply the RAS is so fundamental to their budgetary allocations that disclosure of the algorithm could collapse the entire system; on the other, they want to suggest that it is has no real import, it shouldn't be relied upon and it doesn't allocate anything, so we don't need to know about it.  They can't have it both ways.  But aside from the illogicality of their grounds for refusing disclosure, this belies a more intriguing feature about RAS's obfuscatory and propaganda values.  On the one hand, they are used in support of claims to fairer resource distribution, because they were intended to produce "clear and transparent entitlements", but in reality, they don't produce any kind of enforceable entitlement at all.  My research (yet to be published) suggests that some client groups systematically get less than the 'indicative amount' the RAS spits out, where other groups will tend to get that amount or more.  They are only 'entitlements' when it suits the local authority; they cannot be enforced by somebody who has been awarded less than the RAS indicates they should be.

I've sought a review of Croydon's refusal to disclose the RAS algorithm.  At present they haven't even told me what exemption under the Freedom of Information Act 2000 they're applying, and they're late in responding to my request for a review.  In a way, I'd quite like them to continue to refuse disclosure so we can get a test case to the Information Commissioner's Office...

Hackney: the algorithm is copyrighted...

Hackney's grounds for refusing to disclose the RAS algorithm are rather more interesting:

The information you requested on the mathematical formulae used to covert assessed care needs into indicative budgets is the property of the FACE Recording and Management Systems. So far the company has not made the algorithm available to individuals or organisations because it regards it as distinctive and of commercial value. See In particular see the section in the FACE ‘approach explanation guidance’ entitled: Is FACE Willing to Let People See the Algorithm Used for Calculation?

The document they refer to then says:

FACE wishes to be as open as possible about its work but whilst we have no objection in principle to making the algorithm available to individuals or organisations to date we have not done so. This is for the following reasons:

1. The algorithm is distinctive and of commercial value. It has been consistently been demonstrated to be far more accurate than competitor products and it would clearly not be in FACE’s commercial interest for our competitors to be made aware of the methods used or the details of the algorithm’s functioning. Thus, any sharing of the the algorithm with interested parties would have to be under strict terms of confidentiality.

2. Viewing an algorithm is in itself not a very enlightening process, without a full explanation of how the algorithm was derived and why each step in the algorithm functions as it does. However, even with such explanation it would not be possible for an individual to assess the validity or otherwise of the algorithm since its validity can only be tested across many individuals. Thus, whilst we appreciate that people are naturally curious as to how it works, we are not persuaded that simple publication of the algorithm would in itself prove helpful. 

FACE then go on to say that they'll be producing some documentation explaining their RAS, hosting seminars, academic publications, blah blah blah...  But they won't be sharing the algorithm itself, just the general process by which they arrived at it.

The crucial point surely is that Hackney is entrusting an unspecified algorithm to allocate its budgetary resources, yet apparently it is not allowed to know how it works, and neither are we.  FACE want to be able to flog it to other local authorities, who no doubt would also like to deflect scrutiny from their resource allocation decisions.  We should all just trust it, and not seek to look at its underpinning assumptions and calculations, because... what?

Just to be clear about what we're talking about here, elsewhere where I have been able to look at the underpinnings of RAS algorithms I have found some which systematically ignore certain responses on assessments which could indicate eligible needs for community care services; I have found some which operate extremely harsh deflators for anybody who lives with a carer (with very little sensitivity to what support the carer is willing and able to provide); some are 'universal' whilst others take into account that the unit cost of care services for certain user groups will vary; some make additional provision for people who lack mental capacity and some do not; most RAS's have an upper band beyond which personal budgets must be calculated by traditional means, but these vary enormously; some apply overall deflators so people will get systematically less under a RAS than the value of a traditional budget for a person scoring equivalent points...  You see my point.  To refuse disclosure on copyright and commercial grounds is highly problematic when we consider its function.  This isn't just a swizzy piece of computer software which looks nice, some innocent little interface, it has real implications for people's lives and for the distribution of public monies.  I don't think I can be the only person feeling distinctly uncomfortable with a public authority hiding its decision making processes behind 'copyright' and commercial arguments.  Is this going to be the direction of travel once more public services, and more decisions, get outsourced to commercial providers?

[Update 4 April 2012: I just had a chat with Hackney's RAS team on the phone, who were very open and helpful.  Although they can't disclose the FACE methodology, what they could say was that the FACE system they use is not a points based system.  They also stated that although the FACE system lets you used a deflator (also known as a 'cost abatement multiplier', whereby a RAS takes the typical cost of a package of care for somebody scoring equivalent points on the RAS, then slashes it by a given proportion - usually between 5-20%, to reduce overall expenditure) Hackney had chosen not to use one.  The FACE RAS used by Hackney took into account that the unit cost of care for some groups cost more than others, but they commented that occasionally they used some culturally specific providers for some client groups, and the RAS didn't take this into account - but that the staff knew they'd have to moderate up the indicative amount in these circumstances.  Overall it sounded like one of the more sophisticated systems I've seen, but that doesn't really help with the transparency issues.]

Beyond the democratic and transparency concerns I have, I am unsure how a system like this can be legally accountable.  As Justin Leslie just commented on twitter when I had a little moan about Hackney's refusal, the ruling in Savva v Royal Borough of Kensington and Chelsea states that:

In the present case, the Deputy Judge concluded that fairness required the provision of reasons. He emphasised the consistent theme in the guidance emanating from the Department of Health and the Association of Directors of Social Services – the need for transparency in the decision-making process. In my judgment, he was right to do so. When a local authority converts an established right – the provision of services to meet an assessed eligible need – into a sum of money, the recipient is entitled to be told how the sum has been calculated. It is submitted on behalf of the Council that the imposition of such a duty is excessively onerous for a local authority. By way of illustration, there are some 5000 recipients within the area for which the Council is responsible. It is said that to require the Council to provide each of them with written reasons would be unduly costly. I accept that the burden would not be insignificant but it is what simple fairness requires. If a local authority were entitled to notify a bald figure without any explanation, the recipient would have no means of satisfying himself or herself that it was properly calculated. As the guidance from the Association of Directors of Social Services puts it, explanations of decisions "make it possible for people and families to challenge these decisions". Or, to put it the other way round, an absence of explanations may make it impossible to mount such a challenge, whether by way of complaint or by way of litigation. [20]

What would written reasons for arriving at this calculation look like in the present case?  Would it look like: because the FACE system has assigned a series of unspecified weightings to responses on a questionnaire, and then performed an unspecified calculation, based upon a number of unspecified assumptions about the cost of meeting particular needs, then (may) have applied an unspecified 'deflator' so the local authority can reduce its overall expenditure on adult social care, to arrive at a figure which you can either put up with, or challenge?  In the Court of Appeal ruling in KM v Cambridgeshire the requirement to give reasons was diluted somewhat, and the court ruled that 'a rational link between the needs and the assessed direct payments, but, in our judgment, there does not need to be a finite absolute mathematical link.'  The case has now been heard by the Supreme Court, whose judgment is awaited.  As Nick Armstrong pointed out, writing about the Cambridgeshire appeal, a finding that local authorities must give reasons for their budgetary allocations could have massive implications.

Of course, as almost every local authority I canvassed about their RAS has pointed out, the 'indicative amount' is 'just a starting point' - as if to suggest it doesn't really matter how that starting point is arrived at, because it is not set in stone anyway.  To return to my earlier point - either RAS's have value in helping determine individual budgets or they don't.  If they do, if they are relied upon, then we should know about how they work.  If they don't contribute towards budget setting and resource allocation, it seems reasonable to ask what on earth local authorities are doing wasting their time and money on them?  RAS's increasingly seem to be a policy spandrel, a by-product of personalisation which have more to do with cost cutting and obscuring decision making than empowerment and transparency.  RAS's risk being a pointless and costly ritual that we must all submit to to claim adherence to personalisation, and like a theological ritual, it seems, we are expected to submit to them without question or explanation of how or why they work.