For more than a decade, ministers have argued that digitising public services will unlock a more productive state. Automate. Streamline. Go online. Productivity will follow.
The current evidence says otherwise. Across tax, welfare, justice and health, the same pattern recurs. Digital transformation is increasing spending, embedding complexity and shifting work from government onto citizens and frontline staff. Productivity at the level that matters – the whole system – is scarcely improving at all. And failure to improve productivity in the public sector means rising costs for taxpayers. It also makes it harder to improve productivity, and therefore economic growth and living standards, across the economy as a whole.
This is not a story about isolated programme failures. It is our current delivery model and it is letting us down.
HMRC: Digital as overlay, not simplification
Making Tax Digital was meant to be a flagship programme. Seven years in, parliament’s Public Accounts Committee finds that basic design questions are still unresolved with costs approaching £1.3bn, almost three times higher than first forecast.
More importantly, the committee concludes HMRC has lost sight of its purpose. Taxpayers will spend more and do more to comply. Many will be forced to purchase software. Quarterly reporting will increase workload. HMRC’s own advisory board finds that almost 70% of small businesses expect no benefit from MTD in administering income tax at all.
Meanwhile, HMRC’s internal admin costs keep rising. Legacy systems have to be upgraded and kept alive while new digital tools are developed. The tax system is as complex as ever. The centre gains more data yet struggles to derive meaningful insight from it. As a result, public sector productivity is barely improving.
Welfare: Universal Credit and burden by automation
Universal Credit was sold as a digital-by-default reform. Simpler for claimants. Cheaper for the state.
The National Audit Office could never conclude UC was value for money. Costs rose and delivery slipped. Crucial design features were rebuilt mid-flight. More recent research shows UC’s digital architecture creates three types of administrative burden for claimants: time costs, financial costs and the emotional labour of navigating opaque automated decisions.
UC may be more manageable for the department, but for many claimants it is more work, not less, which is especially egregious given many of those claimants already live challenging lives. The state’s burden has not disappeared. It has merely shifted onto users and frontline staff.
Justice: Courts reform without system productivity
The £1.3bn courts and tribunals reform promised streamlined digital justice. The NAO’s latest review is clear. Some services are not working as expected. Key platforms remain unreliable. Staff routinely operate old and new systems in parallel. That is the opposite of enhancing productivity.
Court backlogs remain high and throughput is down. The digital programme has modernised parts of the system. It has not improved the performance of the system as a whole.
Health: The NHS IT warning that went unheeded
The NHS National Programme for IT remains the most expensive public sector IT failure in UK history. Over £10bn spent, little delivered and years of fragmentation that still affect NHS trusts today.
The deeper failure was not technical. It was strategic. A centralised design that ignored local workflows with an over ambitious scope and a refusal to simplify existing processes before attempting to digitise them. That basic mistake has resurfaced repeatedly since.
Beware the bear traps
Across these systems, the same traps have recurred.
1. Tech first, simplification later
We automate what exists, however complex. In tax, welfare and justice, digitisation has encoded complexity rather than removed it. If you digitise a messy system, you get a faster mess but the mess is still there.
2. Departmental efficiency at the expense of citizen time
If government only measures its own workload, self-service always looks attractive. But small businesses paying for software to comply with MTD, or claimants managing UC journals, are doing work the state once did. Productivity has not risen. It has been transferred.
3. Permanent double running
Legacy systems live on for years because retirement is never properly planned or funded. Courts, UC and HMRC all face prolonged periods operating dual systems. From a whole system perspective, this is ruinous.
4. Benefits cases that ignore human effort.
Departments quantify their own savings but rarely quantify the time, cost and friction imposed on citizens and intermediaries. What is not measured is not managed.
International lessons
Other countries show a different approach is possible. Estonia, for example, redesigned services around a single digital identity and interoperable data infrastructure. Paper processes were systematically removed. Citizen time was treated as a scarce resource. The result is an estimated 2% of GDP in annual time savings.
New Zealand’s Inland Revenue rebuilt its tax system in phased releases, simplifying rules alongside technology. It achieved 99% digital uptake and large admin savings because simplification and capability building went hand in hand.
The message is simple. Technology does not create productivity. System design does.
What a strategic state would do
A strategic state would not treat digital as a shortcut to impossible productivity targets. It would apply four tests before committing resources or political capital.
- Simplify before digitising: No major programme should proceed without measurable reduction in rules, red tape, variants and handoffs. If simplification is deferred, productivity will never arrive.
- Measure whole system productivity: That means calculating the total time and cost for citizens, intermediaries and local partners, not just the department. Treasury and Cabinet Office should require a system productivity metric in every business case.
- Time limited double running and legacy retirement: Every programme should have a dated, costed plan for switching off old systems. Without this discipline, savings remain theoretical.
- Accountability for human impact: Administrative burdens on citizens and front ine staff must be tracked. If burdens rise, the programme is not a success, whatever the internal metrics say.
The choice now
The conclusion is uncomfortable. We have not been digitising the state. We have been digitising complexity. The result is a more expensive system drowning in digital, yet with citizens doing more of the work.
The choice is whether to continue or to reset. A more disciplined, strategic state would use digital to reduce human effort – simplify the machinery of government and improve the productivity of the whole system, not just performance metrics.
That is the standard we should now hold ourselves to.
Patrick Diamond is professor of public policy at Queen Mary University of London and a former head of policy planning in No.10. Vijay K. Luthra is a public service transformation specialist and former civil servant, local government councillor, school governor and NHS NED