The statistical nonsense of the proposed Home Office police funding formula of July 2015 was exposed in my November Opinion piece. The serious irony of the July formula was that it violated the first of the Home Office’s five principles of a good funding model, namely that it should be analytically sound. No one is now defending the proposal but, sadly, that is not because the nonsense has been officially acknowledged.
Requisitioned by the inimitable Keith Vaz MP, chair of the home affairs select committee, a House of Commons debate about the formula took place in March. Excusably unaware of my article’s pressing message, MPs were encouraged to focus on the error for which the minister had to apologise and that had been fortuitously revealed as no more than a wrong key-stroke in the computer input of just one component of one of the five indicators of the formula.
The debate managed to fill three hours with a string of complaints by individual MPs that this or that element of police activity was not represented in any of the indicators, thereby making the formula unfair to the MP’s own constituents. The elusive concept of fairness hung over the chamber, more like a will-o'-the-wisp than a Holy Grail, and MPs did not question the sense of the formula when they asked for their special interest to be included as an additional indicator in the formula. If the revised formula were to oblige them by doing that, the Home Office would be offending its sensible restriction to indicators that do not easily generate perverse incentives.
If the Home Office is indeed determined to use some sort of incentive-free, needs-based measure, the team engaged on the revision should now be trying to find a logical way of weighting the incentive-free indicators to have high correlation with an honest measure of activity-based needs. What are the prospects of the Home Office team being able to do that? Can we even be sure that they will acknowledge the nonsense in the consultation proposal? The Home Affairs committee’s onslaught on the ‘shambles’ of the fortuitous error did recommend that the Royal Statistical Society (RSS) have oversight of any new proposal. It might be expected that such oversight will ensure that the nonsense is exposed and censured, but the portents are not particularly reassuring. Roughly half the RSS membership are government statisticians and, even though the RSS has sacrificed valuable journal space to expose the even costlier shambles of NHS funding formulas (not the work of government statisticians), censure of that bit of policymaking is still awaited.
All parties agree that, if police forces differ in the per capita value of what they do for society, per capita funding should bear some relationship to such differences. The Home Office proposal imagines this can be done, without considering the relative value of individual activities, by a purely empirical technique such as Principal Components Analysis to weight social indicators likely to be correlated with crime or other calls on policing. I would like to think that the Home Office team succumbed to that idea only when exhausted by honest efforts to resolve the real problem. Efforts to bypass difficult value judgements have a long history. The ill-fated Spottiswoode report of 2000 evaded value judgements only by allowing the measure of performance to be the choice of each police force! The Home Office refreshed the evasion in 2003 with ‘spidergrams’ whose complexity met with a wall of disapproval and the Police Federation comment that ‘indicators alone will never capture the whole picture of policing since much of the work that is carried out never gets measured’.
To escape from what may be an infeasible research programme, here is a radical suggestion that I hope the Home Office will consider. It comes out of the long history of NHS formula making. Between 1976 and 1989, the Treasury grant for the NHS was allocated to 14 regional health authorities that were then expected to make good use of their allocations without being told how to sub-divide it. In 1989, the number came down to eight. If a similar trust could be placed in eight restructured police forces with boundaries engineered to embrace populations with roughly the same needs profile, then population size alone could determine the allocation to a broadly acceptable level of fairness – in effect, a vacuously constant per capita formula.
Spreading the areas of London north and south of the Thames outwards to well-defined hinterlands, the Met would make two such forces. Even if the equalisation of social profiles did not achieve the holy grail of perceived fairness, it might be agreed that any attempt to eradicate relatively small and probably contestable inequalities would be too costly and not in the overall public interest. Arguments within forces about allocations of funds would have to be either resolved or adjudicated by a posse drawn from the eight police and crime commissioners wearing the cloak of independence and electable authority. Note that reduction to eight forces for England & Wales would be broadly compatible with the single, apparently trustworthy and formula-free force north of the border for a population of over 5 million.
On March 1 2016, police minister Mike Penning told MPs that the formula revision will be delayed for at least a year, until chief constable Sara Thornton has completed her report of police force ‘capabilities’ for the National Police Chief’s Council. The European Union referendum outcome has now compounded the uncertainty. Can we hope to see the end of the Dark Ages of formula making and a New Dawn of enlightened competence? Now that we know that the next prime minister will be Theresa May, we are surely free to ask her either to apologise for or to disown the OTT prejudgement of the nonsense formula in her foreword to the consultation document? Such questions may never end unless formula making ceases to be the preserve of privileged individuals.
