The unfortunate saga of Universal Credit is a powerful reminder of the repeated failure to successfully integrate policy making with technology. A failure that comes with profound human consequences and suffering, not just a financial cost.

Universal Credit (UC) aimed to deliver a radically better approach to welfare benefits, one that would provide: 

“A dramatically simpler, streamlined system with harmonised eligibility rules, designed to reduce benefit dependency and to be simpler to administer”

Dynamic Benefits. Towards welfare that works

And yet it has frequently dominated the news for all the wrong reasons—becoming synonymous with excessive cost, cover-ups, extended timescales, troubled technology, and iniquitous human consequences.

The original plan was to implement UC over two to three years at a cost of £2.2bn. In reality, by the time it’s fully delivered, it will have taken at least 14 years and a reported £15.8bn.

So what should politicians learn from all this?

David Freud, the former Minister for Welfare Reform, provides a candid account of the difficulties encountered with UC in his book Clashing Agendas: Inside the Welfare Trap, which he also discusses in his Computer Weekly article Universal Credit: What went wrong, and what we learned.

In this piece, I explore some of the backstory of tax and welfare modernisation. It provides a timely reminder of the need for improved integration between policy and technology throughout their entire life cycle. It draws on meetings, discussions, documents and Parliamentary, National Audit Office and other evidence.

It’s longer than I originally intended. For those too busy to read it in full, here’s a brief up-front summary of some of its conclusions.

Integrate policy and technology

The problems encountered with Universal Credit illustrate the need to design and implement technology and policy more effectively together. There should not be a sequential process of top-down “Policy Decision” and then “Technology Implementation”.

Policy is not a once-off, “fire and forget” process, but needs constant monitoring and adjustment. To do that, policymakers and officials need to take advantage of the processes of continuous improvement typical of successful digital organisations. 

Work in the open

One of Whitehall’s long-standing problems has been its reluctance to work in the open and to admit to, and learn from, its mistakes. The prevalent “fear of blame / headlines in the Daily Mail” culture needs to be replaced by one of open sharing, learning and improvement.

All non-secret government programmes should work in the open. One reason politicians, policymakers and technologists keep making the same mistakes is​​ that it’s impossible to learn and improve if they can’t be honest and open about what’s worked—and what hasn’t.

Treasure and nurture creative officials

In the early 2000’s, a small team of Whitehall officials worked with the UK national banking and payments operator to develop a range of innovative ideas for the improvement of income tax—ideas from which the concept of dynamic benefits was born. They created the space to think differently about taxation and how to implement it more effectively in the twenty-first century.

This sort of fundamental rethink and redesign of the way policy and technology interact needs to be at the core of political thinking and “digital transformation”, not automating the status quo or simply improving the design of front-end screens.

For this to happen, it will require not only policymakers with a vision and commitment to modernise the way government operates, but also digitally savvy officials with the ability to understand how to do it—and the authority to see it through. Government should treasure and nurture the creative officials who step outside the group-think of performative digital theatre, and who generate innovative, viable ways of redesigning and replatforming government.

Without cultivating the right people, ideas and expertise, political parties will continue to struggle to deliver the promises that brought them into office—and technology will impede rather than assist them.

PAYE improvement, dynamic benefits and Universal Credit

Three of the early papers that led to Universal Credit 2007-2009

Universal Credit is often associated with the coalition government of 2010 and subsequent Conservative administrations. However, its origins lie in plans to improve welfare under the earlier Labour government, with the publication of David Freud’s 2007 independent report Reducing dependency, increasing opportunity: options for the future of welfare to work. It recognised that:

“The Government’s aspiration to achieve an employment rate of 80% is extremely challenging. Its achievement would establish the UK as a world leader in employment. It would also produce many other substantial benefits in helping to deliver other social goals, including, most importantly, that of reducing child poverty.”

In his book, Freud recalls how all three of the major parties in Parliament endorsed his paper. Significantly, it foresaw the need to move to a single system of benefits:

“There is a strong case for moving towards a single system of working age benefits, ideally a single benefit, in order to better support the Government’s ambition of work for those who can and support for those who cannot. A range of international evidence suggests that complexity in the benefit system acts as a disincentive to entering work, and that badly designed systems create unemployment and/ or poverty traps.”

In his foreword to the follow-up white paper, Prime Minister Gordon Brown wrote:

“Too many of our communities still bear the scars of previous downturns, never having recovered from the scandal of millions pushed into a benefits system that too often trapped its recipients rather than supporting them back into work … This White Paper marks the next stage in our welfare reforms – moving towards a system that offers more support but that expects more in return.”

The later genesis of what was to become Universal Credit was the 2009 Centre for Social Justice’s report Dynamic Benefits. Towards welfare that works. It recognised the plans were ambitious:

“We do not underestimate the nature of this change, and would expect that it would be best to allow two to three years to build the computer systems.”

Dynamic Benefits. Towards welfare that works (my emphasis)

The aim was to deliver:

“Tax and benefits withdrawal integrated into a single system, “PAYE+”, making it more accurate and responsive to changes in earnings, with reduced risks involved in returning to work”

Dynamic Benefits. Towards welfare that works

These ambitious ideas for welfare reform based on “PAYE+” depended for their success on earlier plans for the radical overhaul Pay-As-You-Earn (PAYE)—an income tax system introduced in 1944 during the second world war and, although now automated, left largely untouched ever since.

The backstory

Select “highlights” from the Universal Credit story

Plans to modernise PAYE date back to at least 2006. Known internally as PIG (PAYE improvement) and externally as PAYE2, the aim was to rationalise, integrate and accelerate the delivery of welfare and the collection of National Insurance contributions and taxes. It would improve services for taxpayers and welfare recipients alike, while providing operational efficiencies and lower costs for government and business.

“… the labour market today is significantly more complex and the modern worker is more flexible and mobile. For example:

— Higher amount[s] of taxpayers hold multiple sources of income (the number of second jobs increased by 68% between 1984 and 2001);

— Casual employment is more common;

— Average job time is shorter (20% last less than one year and 5% less than 3 months); and

— People move job more frequently (13 million people moved job in 2006).

PAYE at the Crossroads. All Party Parliamentary Taxation Group, July 2012.

At the heart of PAYE improvement was the idea of government digital platforms hosted within the UK national banking and payments system operated by VocaLink. By 2011, this payments infrastructure was already handling over 90 million transactions on a peak day, including more than 95% of salaries and 98% of benefit payments. Re-use of this existing, secure national infrastructure would help provide accurate individual and household income data, accelerating reform of welfare and taxation and propelling both into the twenty-first century.

The new service would hold full details of what everyone was about to be paid and their correct tax position at that moment in time. Employers could use it to get an accurate and real time calculation of how much tax to deduct from an employee rather than relying on the creaking system of annual tax codes.

Dynamic benefits were to provide similar improvements for welfare. Their success relied on being able to access the full details of someone’s income and tax position. A real time benefits calculation and resulting payment could then be made that accurately reflected the recipient’s situation.

As both tax and welfare rules became integrated into the central calculation platform, a single set of deductions and allowances could be determined, with the net result paid into the recipient’s bank account.

In its final version, an employer wouldn’t even need to make any deductions: they would pay salaries gross. As payments crossed the banking network, a sub-second, real time calculation would take place, with tax deductions and welfare allowances determined. The resulting net payment would go into the recipient’s bank account, and any taxes deducted into HMRC.

Simplified overview of the proposal for integrated tax and welfare using the UK’s national payments infrastructure

Integrating welfare and taxation provided many potential improvements, including:

  • Dynamic assessment of earnings across multiple employments, with the deduction of applicable taxes in real time rather than relying on the annual tax coding process of the mid-twentieth century 
  • A significantly reduced burden on employers, who would no longer be responsible for the often time-consuming process of determining an employee’s personal tax position
  • The eradication of prohibitive withdrawal rates applied to anyone moving from unemployment benefits to employee, wage-earning status
  • A more cost-effective determination of household income
  • The elimination of conflicting allowances and deductions (taxation of benefits, tax credits, and so on)
  • The potential to move away from the “pull” model, where individuals needed to fill in paper or online forms to claim and prove entitlement to welfare, to a data-enabled “push” model, where the system could determine entitlement in near real time and ensure welfare reached those eligible to receive it

It would end the age-old process whereby one part of government (HMRC) takes money away from someone and another (DWP), belatedly, often hands some of it back: no-one would design government organisations and their services in this dysfunctional and often frustrating, frictional and humiliating way if they were being created now. The original aspiration of PAYE2 was to provide a managed transition from old world to new, delivering a digital transformation worthy of that much overused phrase.

This fresh approach had significant political upside: it would enable political parties to implement their taxation and welfare manifesto commitments quickly and effectively. The system would provide one place where changes could be implemented to welfare, or changes to NIC and tax rates. In future, any government could rapidly adopt a different threshold or otherwise adjust the tax and welfare rules in near real time to meet their own socio-economic goals. This ability to automate policy and support dynamic changes by policymakers was integral to the original approach.

The Centre for Social Justice report, for example, proposed to reduce the rate at which benefits are withdrawn to a single rate of 55% of post-tax earnings. However, the rate initially applied by the government was a much controversial 65%. It’s now 63%, effectively acting as a 63% marginal tax rate (or 75% when national insurance and other taxes are included). A well-designed, integrated system would have enabled a government not only to flexibly adjust such thresholds, but also to model various options—their costs and benefits, and their impact on taxpayers and welfare recipients—before deciding which option to implement.

In 2010, an early pilot system, developed using policy automation software already used in government and elsewhere, demonstrated how dynamic welfare could work in practice: 

“[The program] translated documents into executable business rules, which meant it could work out people’s entitlement to Universal Credit from our policy documents and legislation. I had tested some basic propositions and the programme’s ability to respond accurately had left me deeply impressed.”

Clashing Agendas. Inside the Welfare Trap. David Freud, 2021.

Along with detailed design and implementation proposals, this pilot helped provide assurance that the estimated two to three years timescale to develop and deliver both the policies and the technology was viable, with live service anticipated for 2013. However, the re-use of existing commercial software was later to be discarded and replaced with bespoke development work.

Ideas meet reality

As the ideas developed, PAYE2 became ARTE (Accurate and Real-Time Earnings), and dynamic benefits became Universal Credit.

ARTE was intended to be a shared government system integrated into the national payments infrastructure and making use of the ISO20022 standard. It would hold the taxable payment history for citizens for the current and previous tax year, enabling DWP and HMRC to determine tax allowances and welfare entitlements. By re-using the existing UK payments infrastructure, ARTE would provide timely individual income data and create an accurate view of household income.

Simplified overview of Accurate and Real Time Earnings (ARTE)

Account information would build up over time as employers moved to the new service. ARTE would contain a limited amount of information relating to households, taxpayers and employing organisations. A set of interfaces would enable HMRC and DWP to access and query the ARTE database at an individual and household level to ensure tax and benefit decisions used the most up-to-date information.

However, this is not what happened. Instead of shared platforms integrated into the UK’s national payment infrastructure, the programme fragmented into multiple Whitehall technology projects. The result was a series of parallel departmental contracts, designs, systems and databases built by various suppliers and teams, many of them new to the concepts of dynamic policy, data-enabled services, real time payments, welfare and taxation. 

“At the heart of UC’s troubles lay the capability of the department [DWP] in IT. It became evident that the commissioning out of IT by government departments a decade and more earlier had effectively removed any direct knowledge of how to build systems, or even monitor that the contractors were building those systems properly … At the same time, the first attempt to use new agile development techniques fell apart because the approach used – building up individual story lines – was highly inefficient.”

Universal Credit: What went wrong, and what we learned. David Freud, Computer Weekly, 25 June 2021.

The significant decision was taken not to integrate ARTE functionality within the existing payments infrastructure. This was despite the All Party Parliamentary Group on Taxation observing in 2011 that:

… it may make sense for HMRC to utilise information already held by existing third parties rather than to build and operate its own in-house infrastructure.

PAYE Review and Recommendations. All Party Parliamentary Taxation Group, 2011.

Not long afterwards, Government Digital Service (GDS) guidance also recommended that:

“A move to platforms does not mean that government has to develop everything in-house: many of government’s needs can be met by existing cost-efficient utility services … Wherever appropriate, the government should use existing external platforms, such as payments services (ranging from third party merchant acquirer services to the UK’s national payments infrastructure). Deciding to develop platforms in-house will happen only where that is the best way to meet users’ needs in the most flexible and cost-effective way.”

UK Government Service Design Manual, 2013. Deleted content.

ARTE was descoped into Real Time Information (RTI), which was further broken down into an “interim” and “strategic” solution. In 2012, interim RTI was described as:

“a stepping-stone, not the final destination. HMRC both can and should go a lot further with the modernisation of PAYE.”

PAYE at the Crossroads. All Party Parliamentary Taxation Group, July 2012.

The “interim” model for Real Time Information (RTI)

David Freud describes the situation in his book:

“The Vocalink PAYE2 proposal had been to track wage payments as they went through the banking system, producing an automated real-time data feed to HMRC. The interim system … had a very different approach. It relied on employers forwarding the information about wage payments to HMRC, as soon as they were made. HMRC would then make that information available to DWP. In practice it meant an exercise to get the providers to add an extra button to their payroll software which employers could use to transmit the information about payments immediately after the bank transfer instructions had been sent. It also meant abandoning Vocalink as a supplier, despite their pivotal role in developing the conceptual base for Universal Credit. My efforts in the following years to move to their real-time approach took me down several byways of the banking system but I was unable to resuscitate PAYE2.”

It was significant that the new, in-house systems would no longer have accurate tax and payment information across concurrent employments in real time. Decisions being made at a technical and delivery level were removing functionality essential to achieve the original policy objective. While HMRC has made good progress with the more limited RTI programme, it currently provides only a subset of its original intent, and has been the subject of insider criticism.

This drift away from a unified, shared and real time system for tax and welfare integrated into the UK’s highly resilient national payments system arguably laid the seeds for many of Universal Credit’s problems. Instead of a single comprehensive database of earnings, the new HMRC database contained only partial information—information that was copied and duplicated into another new database held in DWP to provide the basis for Universal Credit.

Far from being ‘real time’, these separate departmental databases were developed using old-fashioned, twentieth-century batch transfers. The chance to radically improve welfare for those reliant upon prompt and accurate payments was rapidly slipping away. It’s little wonder that the Major Projects Authority Project Assessment Review of February 2013 awarded Universal Credit an unqualified and embarrassing RED. 

Reality bites

Instead of taking only two to three years, Universal Credit is unlikely to be fully live until at least September 2024. That’s 14 years since the planning and implementation of Universal Credit started under the coalition government of 2010. And it’s 15 years since the Centre for Social Justice report, 17 years since David Freud’s independent report for the last Labour government, and 18 years since the proposals to improve PAYE.

These extended timescales should alarm any political party aiming to make timely policy changes and improvements, as should their associated costs—financial and human. These are pre digital, paper-world timescales, not the sort of agile, nimble processes of continuous improvement typical of digital organisations. 

They illustrate just how far away government remains from where it needs to be with digital, data and technology. Yet with Universal Credit, this failure is not through any lack of political vision or ambition. David Freud, then Minister for Welfare Reform and initially responsible for bringing the project in recently observed

“We had set out on the project to reform the UK welfare system in 2010. The existing system was a mess – a bundle of often contradictory benefits assembled over the years that categorised and trapped people in the system and in poverty. With Universal Credit (UC), we wanted to sweep this all away and replace it with a system that allowed people flexibility to change their situation in a low-risk way.”

The reality has, however, been markedly different. As Stephen Timms MP, Chair of the Work and Pensions Select Committee, has commented:

“The irony … is that despite hundreds of millions of pounds of investment in digital technology, an initial benefit payment is now slower than at any point in the history of the welfare state. Ever since the 1940s we have had a social security system that has been capable of delivering the first regular benefit payment to people within a few days … It seems to me this is not a system that is fit for purpose if it cannot do what the system has done ever since the 1940s.”

Lessons to be learned

So what went wrong?

The best modern organisations exploit technology to provide real-time and historic data and insights to help better evidence, analyse, inform and update their decisions. They use it to create and operate more effective organisations and services, slipstreaming technical improvements into their services multiple times a day as and when needed.

In the public sector, however, technology is regularly relegated to the delivery of a top-down policy decision rather than using it to help continuously inform and improve policy. Often it has simply enabled “disaster, faster”.

Perhaps this is no surprise when “digital” training courses for senior officials and politicians focus on topics such as “agile” and “scrums”, reinforcing the view that it’s solely a mechanism of policy delivery and administration rather than integral to policy design. Yet an agile approach to technology won’t transform Whitehall unless the waterfall approach to policy and decision-making is overhauled too.

The current “training” needs to be replaced with meaningful education about digital business models instead. Doing so will help cultivate expertise, experience and confidence within the public sector to use digital, data and technology to inform policy making, and to rethink and re-platform government for the 21st century.

So what should politicians and policymakers learn from this “flagship” programme? Here are a few thoughts to add to those David Freud sets out in his book.

Integrate policy and technology

The problems encountered with Universal Credit illustrate the need to ensure technology and policy are designed and implemented effectively together. They should not be a sequential process of top-down “Policy Decision” and then “Technology Implementation”. 

Policy is not a once-off, “fire and forget” process, but needs to be constantly monitored and adjusted. To do that, policymakers and officials need to take advantage of the processes of continuous improvement typical of successful digital organisations. 

Technology decisions made during the development of PAYE2 and hence Universal Credit had serious policy implications, breaking some of the elements essential to “dynamic benefits”. David Freud recommends a reversal of “a policy-led culture to one reliant on operational feedback”.

The processes of continuous feedback and improvement that data and technology can provide, enables policy to be refined and improved as it is implemented, with modern systems designed to be easy to change in near real time. The challenge here is neither new nor primarily one of technology, but of needing:

“… to work in new ways if the opportunities to improve public services are to be seized. It will need to re-invent how it works through stronger leadership from the top, clearer and more powerful incentives to change, radical shifts in arrangements for working across boundaries and a cultural change to support innovation.”

Electronic Government Services for the 21st Century. A Performance and Innovation Unit Report, Cabinet Office, September 2000.

These words ring as true now as when they were published back in 2000. No mainstream political party has currently set out a well-developed, comprehensive political vision of how to use digital, data and technology to reform and re-platform government for the 21st Century. Even when a vision exists in discrete policy areas, such as tax and welfare, implementation has fallen well short of the original policy intent, and often undermined it.

Work in the open

One of Whitehall’s long-standing problems has been its reluctance to work in the open and to admit to, and learn from, its mistakes. As Computer Weekly reported in 2018, in Universal Credit: How did it go so wrong?, DWP repeatedly masked serious governance and technical problems until being forced to reveal them only after a 4 year long Freedom of Information battle. The prevalent “fear of blame / headlines in the Daily Mail” culture needs to be replaced by one of open sharing, learning and improvement.

Universal Credit is a significant public programme involving large amounts of public funding. Welfare goes to some of the most needy members of society. Yet unless changes are made to this endemic culture of secrecy and its associated lack of accountability, we can expect policy and technical failures, and their associated massive cost overruns, to be repeated indefinitely.

All non-secret government programmes should work in the open. One reason politicians, policymakers and technologists keep making the same mistakes is​​ that it’s impossible to learn and improve if they can’t be honest and open about what’s worked—and what hasn’t.

Treasure and nurture creative officials

In the early 2000’s, a small team of Whitehall officials working with the operator of the UK national banking and payments network developed a range of innovative ideas for the improvement of tax—ideas on which the concept of dynamic benefits were founded. They created the space to think differently about taxation and how to implement it more effectively in the twenty-first century.

Much of the work labelled “digital transformation”, however, continues to follow the pattern of the last three decades, frequently automating and tweaking what already exists within existing policy and organisational silos. The day-to-day pressure of “keeping the lights on” of current critical services understandably dominates short-term thinking and behaviours, but often to the exclusion of modernising and improving the way government works.

So what might true “digital transformation” look like? Consider the two figures below. Instead of simply automating old processes, they rethink and reinvent them in an entirely digital way.

The first shows how the PAYE system has traditionally worked, a transaction-intensive approach that simply mirrors and automates older paper-based processes. 

PAYE as mapped before PAYE2

The second shows how the entire system could be rethought, streamlined and simplified if the full PAYE2 vision, including a central calculation platform, were eventually implemented.

PAYE as imagined after PAYE2

It’s this fundamental rethink and redesign of the way policy and technology interact that needs to be at the core of political thinking and “digital transformation”, not automation of the status quo or simply improving the design of front-end screens.

For this to happen, it will require not only policymakers with a vision and commitment to modernise the way government operates, but also digitally savvy officials with the ability to understand how to do it and the authority to see it through. Government should treasure and nurture the creative officials who step outside the group-think of performative digital theatre, and who generate innovative ways of redesigning and replatforming government. 

Without cultivating the right people, ideas and expertise, political parties will continue to struggle to deliver the promises that brought them into office—and technology will impede rather than assist them.

Select further reading

What’s striking in all of the Universal Credit reviews and assessments made public is that none of them appear to reference the original approach—to create a 21st century infrastructure to reform tax and welfare integrated into the UK’s national payments system. Yet this shared central infrastructure was central to the ideas and implementation plan for PAYE modernisation and dynamic benefits. This backstory is essential to understanding why things went so badly wrong.

Only in David Freud’s recent book do we find many of the important details missing from the many reviews, audits and enquiries that have taken place. The damaging schism he exposes between policy and technology—and between the egos of various “old guard” and “digerati” officials—prevented the full modernisation of tax and welfare. That failure to seize the chance to modernise matters: it’s a significant cause of much of the turbulence and human suffering that have haunted Universal Credit ever since.

As the operators of the UK’s payments infrastructure, VocaLink, commented as long ago as 2012:

“The fundamental concept … was the reuse of existing infrastructure, not the development from scratch of new IT systems. Although HMRC … has, in VocaLink’s view, departed from this original concept in some important respects, it has, in great measure, avoided the temptation to reinvent the wheel and build unnecessary new systems. The challenge therefore is not to make something new but to take what is already available (and working) and ensure that the right components are deployed in the right way at the right time so that a new service (Universal Credit) can be built from pre-existing components.

Secondly, and related to the first point, the delivery of Universal Credit is not solely in the gift of DWP. In order for UC to be delivered successfully, and on schedule, DWP needs to work closely with a number of other major stakeholders, including, of course, HMRC, and VocaLink, as operators of the interbank payments systems, but also the rest of the banking infrastructure, the banks themselves and agencies such as the Post Office (for example in relation to Post Office Card Accounts). The challenge for DWP, as VocaLink sees it, is therefore to act as an enabler, ensuring that the disparate elements required to come together for UC to work are assembled in the right way and at the right time. This is undoubtedly a significant challenge but it is not the same challenge as building a new IT system from scratch …”

VocaLink evidence to the Work and Pensions Committee. Universal Credit implementation: meeting the needs of vulnerable claimants. 14 August 2012

In his book, Freud observes that:

“The time horizons of a major programme of change fit remarkably poorly with the political cycle. It takes many years to effect major change, yet typically neither politicians nor civil servants are allowed the continuity required”

To address the damaging policy / technology divide, Freud floats the idea of an “Implementation Minister” based in the House of Lords, who would be:

“… responsible for implementation of the project within the overall ministerial team and would work for the Secretary of State of his Department. He is also likely to have direct contact with senior members of the Cabinet, including the Prime Minister and Chancellor. At the same time, he will have to ensure that his team is built in a way that can deliver the project but integrates effectively with the rest of the Department.

The dividing line between detailed policy design and implementation would be rubbed out with this appointment.”

These last words are also from Lord Freud. They’re a good place to finish—I think they go to the heart of what Universal Credit, and true “digital transformation”, needs to deliver:

“We’ve been looking at this as a technology issue. It is much more than that, [it’s] a major cultural change in the relationship between the state and the people it needs to support.”

NAO reports

Universal Credit: early progress. 5 September 2013 

Universal Credit: progress update 26 November 2014

Rolling out Universal Credit 15 June 2018

Universal Credit: getting to first payment 10 July 2020

Project Assessment Reviews (PARs)

Major Projects Authority Project Assessment Review (PAR) 10 June 2013

Major Projects Authority Project Assessment Review (PAR) 3 March 2014

Parliament

VocaLink evidence to the Work and Pensions Committee. White Paper on Universal Credit. December 2010

VocaLink evidence to the Work and Pensions Committee. Universal Credit implementation: meeting the needs of vulnerable claimants. 14 August 2012. 

House of Commons Work and Pensions Committee Universal Credit Project Assessment Reviews 24 January 2018

Articles and books

Universal Credit IT: What we know; what we don’t know. Bryan Glick. Computer Weekly. 12 December 2013

Released documents show scale of Universal Credit problems before ‘reset’. Lis Evenstad. 13 April 2016

Universal Credit: How did it go so wrong? Lis Evenstad. Computer Weekly. 12 March 2018

Clashing Agendas: Inside the Welfare Trap. David Freud. June 2021.

Universal Credit: What went wrong, and what we learned. David Freud. Computer Weekly. 25 June 2021

————————————————————————————

This an interim study for a book exploring the interdependencies between policy, technology and society. It will (hopefully) be published sometime late in 2022.

Original source – new tech observations from a UK perspective (ntouk)

Comments closed