By Colin Talbot

28 Nov 2019

At a time when people are more aware of how they can help their local community, Colin Talbot asks if government benefit and taxes could be more effectively delivered with greater local control

Liberals and others have long argued for decentralization of the UKs overcentralized state. Most of the ideas involved either devolution to Scotland, Wales, Northern Ireland and English regions and/or decentralization of Whitehall functions to local government.

Strangely, the decentralisers have rarely, if ever, considered the three biggest functions of central Government – benefits, prisons and tax collection. Sixty percent of civil servants nationally are involved in “operational delivery” and the vast majority of those are in these three functions. Yet for some reason they are largely ignored in discussions about decentralization.

Of these three functions only one – prisons – is devolved. Scotland and Northern Ireland (but not Wales) have their own prison services.


The other two – benefits and tax – are completely centralised with some small exceptions. Local government taxes being collected by local authorities. Until recently local government was also responsible for paying out Housing Benefit but that is now also being centralized through integration into Universal Credit. In Scotland, powers have recently been devolved to two new agencies - Revenue Scotland (formed in 2015) and Social Security Scotland (2018) - to run some localised tax collection and benefit payments, and the Welsh Revenue Authority was similarly established to collect four taxes from 2017. But the main taxes – income, corporation, VAT – and big benefits – universal credit and pensions – remain controlled and administered by HMRC and DWP in London.

Both of these are recent and limited reforms, and the big systems have been centralised for so long that it now seems to be the natural order of things. But why?

One argument for centralisation is that the making of policy on benefits, prisons and tax needs to be combined with “operational delivery” to ensure national standards and avoid the dreaded “post code lotteries”.

Even with the bounds of the UK’s ramshackle state this argument is demonstrably nonsense. There are numerous policy areas – health, social care, education, policing, fire, etc – where national (or English) policies are set and main funding determined centrally but operational delivery is localised in one form or another.

Internationally there are numerous examples of different configurations of benefits, prisons and tax administration.

In the USA and many other countries prisons, for example, are organised at several levels of government from local, through state and up to federal levels.

In Denmark personal taxation and benefits are organized through local government although policies are still set centrally.

So, is it simple institutional inertia and habit that stops UK reformers even thinking about transferring these functions to local government?

“The prison service in England and Wales was too big, too centralised and too clumsy to deal adequately with prisons as a function”

Let’s consider prisons first. At the moment we have one prison service (about to be more closely linked to probation) covering England and Wales.

The fact we have two, much smaller, services covering Scotland and Northern Ireland barely registers on the radar of the Westminster bubble. This partly because it’s not a problem.

There have been no scandals or controversies over the Scottish and NI services because they are smaller and devolved. If anything there have been some small benefits in having three systems with slightly different approaches to compare. (In the USA prisons policies are subject to much, beneficial, comparative analysis across the 50 states).

In the English and Welsh prisons system we have had plenty of incidents and controversies. One such occurred in 1995 when Derek Lewis, the director general of the Prison Service, was dismissed by home secretary Michael Howard after a couple of high-profile escapes from maximum security prisons.

In the spring of 1997 Howard established a ‘Prison Service Review’ to examine the organization and management of the service. I was one of two external members of the Review Steering Committee (along with Sir Michael Heron, then-chair of the Post Office).

The main lesson I took away from the review was that the prison service in England and Wales was too big, too centralised and too clumsy to deal adequately with prisons as a function. This view was reinforced a few years later when I had chance to study the Swedish, Finnish and Dutch prisons systems with my late colleague Christopher Pollitt and also to do some work Corrections Canada, their federal system.

The vast majority of prisoners in most systems are young men doing short-term sentences. All the evidence and expert opinion I have heard from across these various systems is that the best way to handle this population is in local prisons, close to where they lived, with close ties to external services – especially probation and education.

Yet there is a history in the English and Welsh prison service of constant pressure to build bigger prisons, often in remote (and therefore cheap) locations, and to under-provide prison places so that prisoners are frequently shuffled around, often ending up hundreds of miles from their home areas.

The only way to fix this would be to massively decentralise (most) prisons and have them run locally, in close conjunction with police, probation, education, health and other local services.

Taxes and Social Security

Taxes in England were, until the end of the 17th century, collected locally. Even as systems of national taxation through the Inland Revenue gradually came in from the late 1600s, taxes were organized through more localised ‘collections’, a system that carried right through to the current era. Customs and excise followed a similar pattern as it developed with national policymaking and organisation combined with more localised ‘collectors’.

When the then Conservative government launched the ‘Next Steps’ programme of ‘agencification’ of the civil service in the 1990s, both Inland Revenue and HM Customs were reformed and said to be “operating on agency lines”. Agencies were supposed to mean, in this context, separating policymaking from ‘operational delivery’ (although it was never that simple in practice).

So inside IR and Customs the regional ‘collections’ – around thirty in each – were designated as quasi-agencies, with all the paraphernalia of agencies including performance targets, framework documents setting out their objectives, and some managerial decentralisation.

So the principle of centralised policy-setting and more decentralised administration of tax collection is not unheard of, even in the UK.

Having taxes and benefits locally administered would substantially expand local government’s structures”

There are several potential advantages to localising the delivery of these national (UK or England & Wales only) services.

The first is about money. Having – say – local government collect income taxes, national insurance and maybe VAT and pay out benefits including pensions and Universal Credit would create a whole new localised treasury function with substantial cash flows. As all businesses know, having large cash flows confers all sorts of benefits.

The second advantage is around organisation. Having taxes and benefits locally administered would substantially expand local government’s structures, which would bring benefits in terms of increased attractiveness to staff and managers, possible efficiency savings from back-office integrations and capital expenditure, especially on buildings.

The latter would also provide the possibility of creating local “one stop shops” for services where people could go to sort out a range of government related issues.

The third advantage is around ‘authority’ – local government would really be much more of a local government, creating a much closer psychological link between citizen and state.

And this relates closely to the fourth area of advantage: information. Local government would have much greater and more integrated aggregate data about its area, allowing it much better policy and planning insight.

There is some evidence from Scandinavian countries that localised tax administration improves collection rates because of informational advantages of local services (they know their areas) and better trust between citizen and (local) government.

These are the benefits of challenging the centralised mind-set that these big services must be controlled and delivered centrally. Ultimately, all services are local.

Read the most recent articles written by Colin Talbot - Productivity puzzle corner: The Efficiency Wars redux

Share this page