Antonia Romeo sets out civil service performance reforms

Cabinet secretary Antonia Romeo has set out plans to change how Whitehall assesses and rewards its staff, arguing that better performance is necessary when public finances leave limited room for manoeuvre. Speaking at the London School of Economics, she proposed stronger incentives for delivery and expertise, alongside more effective action when officials fall short.

According to the Guardian’s account of the speech, Romeo said the civil service needed to confront criticism that it was excessively cautious, responded to events rather than planning ahead, and struggled to work across organisational boundaries. These were her assessment of Whitehall’s weaknesses, rather than findings from a published performance study presented in the report.

The proposed approach would recognise not only results but also specialist knowledge and persistence in completing work. Romeo’s emphasis on remaining with a task adds a retention objective to the performance agenda: she wants the reward system to value officials who see assignments through, rather than treating movement between posts as the main route to advancement.

Managers would receive training and stronger processes to help them recognise staff making the greatest contribution and deal with underperformance. Romeo described the intended response to poor performance as “decisive action”. The report does not specify what measures that would involve, or set out revised appraisal criteria, disciplinary procedures or dismissal rules.

Andy Burnham and his deputy, Louise Haigh, attended the lecture. Their presence gave the proposals a political audience as well as an administrative one, although the detailed case for changing staff management was made by Romeo, the country’s most senior civil servant.

Romeo linked the need for productivity improvements to the “tight fiscal context”. Ministers face the task of accommodating higher defence spending in future while protecting public-service budgets. Her argument is that changing how the civil service works is essential in those circumstances; the report provides no estimate of how much money the proposed reforms might save.

Changes to increase bonuses for senior civil servants judged to be performing well have already been made. Romeo presented those changes as an initial step rather than the completed reform. The source gives neither the size of the increases nor the number of officials eligible, so it does not establish their overall cost or effect on performance.

A separate problem concerns the incentives created by career progression. Some officials move into different jobs at a similar level to obtain higher pay, rather than being rewarded for continuing in their existing roles. Romeo said she wanted to reduce that turnover. The account does not quantify how frequently it happens or identify which departments are most affected.

Her promise to “rewire” the civil service also follows language used under Keir Starmer’s government. That places the announcement within an existing reform agenda rather than establishing it as an entirely new ambition. The report does not compare the proposed arrangements with earlier commitments or assess which previous measures have been implemented.

Romeo also placed the reforms against uncertainty arising from artificial intelligence and international political tensions. Those pressures formed part of her argument for organisational change, alongside the immediate budget constraint. The speech therefore addressed both the management of individual officials and Whitehall’s ability to adapt to wider changes.

Romeo’s case for tackling underperformance was that it affects colleagues as well as the work of the individual concerned. She argued that tolerating it discourages stronger performers and weakens their willingness to contribute additional effort. This is the management rationale behind her proposals, although the source does not include evidence measuring that effect within the civil service.

On AI, she argued that reducing bureaucracy could leave officials with more time for creative work, interaction and decisions requiring human judgement. Her stated aim was to make public services “more human”. That presents automation as a way to change the content of civil servants’ work, rather than simply as a means of reducing the workforce.

Some civil servants nevertheless worry that AI will be used to replace officials, the Guardian reported. No named staff representative or union response is included. The available account therefore establishes that concern exists, but not its extent, and does not describe a workforce agreement or employment guarantee accompanying the technology proposals.

Romeo cited screening for certain cancers and identifying the risk of homelessness as areas where technology could support earlier intervention and more tailored services. She qualified that argument by referring to the need for controls and safeguards. These were examples of potential applications; the report does not provide results from particular schemes or evidence of their effectiveness.

After the lecture, a government spokesperson confirmed Burnham’s support and connected the reforms to his intention to transfer power from Whitehall to communities and regions. The spokesperson also linked responsible AI use and less bureaucracy to better services, restored trust and growth. Those are the government’s intended outcomes, not demonstrated results of the announced changes.

Romeo is overseeing a review of the civil service’s organisation, with findings expected in the spring. An update to the civil service code is also planned. The account does not identify the proposed code changes, so it would be premature to infer alterations to particular duties or standards.

The organisational review will consider whether existing departmental arrangements are suitable for work that spans different areas of government. Its remit includes examining current structures, but the report names no departments for merger, abolition or redistribution of responsibilities. A review of the machinery of government is not, at this stage, a settled restructuring plan.

AI and what Romeo called “devolution by default” will also feature in that examination. This brings questions about technology and the location of decision-making into the same review as departmental organisation. No specific powers, regional bodies or transfer timetable are identified in the reported proposals.

Several practical details of the performance reforms remain unspecified in the source: the implementation timetable, how expertise and delivery would be assessed, and how rewards for staying in a role would operate. Nor does it give a productivity target against which the programme could be judged. The announcement establishes the intended direction, while the operational arrangements remain to be set out.