Who regulates the regulation?

Why the Regulatory Policy Committee exists

Westminster is full of famous institutions which are much less important than they would like you to think they are, and only a handful of obscure ones which are much more important than anyone realises. The Regulatory Policy Committee (RPC) belongs in the second category.

The RPC might be the most important quango (Quasi-Autonomous Non-Governmental Organisation) that most people have never heard of. It’s the independent watchdog for regulatory analysis – it tells the Government whether the evidence behind new regulation adds up.

Quangos are a popular target for think tanks and reforming governments alike, often with good reason. In many cases, their independence from ministers has become an excuse for inefficiency and waste. In some cases it has resulted in some public bodies acting to further their own objectives instead of those of elected politicians.

The Prime Minister has criticised a “cottage industry of checkers and blockers” in which quangos are “using taxpayer money to stop the government delivering on taxpayer priorities”. It was in the same speech that he committed to abolish NHS England and bring the health service back under the direct control of the Government (in line with Re:State’s recommendations from 2024). And the Government plan to go further by reviewing all public bodies, as we called for last year.

But it’s just as important to recognise when quangos are a good thing, and their independence is playing a crucial role in getting good policy outcomes. Better regulation is a good example.

Better regulation needs better evidence

The PM said last year that he believed “deregulation is now essential” and Labour’s Regulation Action Plan commits them to cutting 25 per cent of compliance costs from regulation over the course of the Parliament. Peter Kyle said he wants regulation to be “well designed and carefully implemented” so it can promote growth – the RPC are the people doing exactly that.

Established in 2009 as a non-departmental public body, the Committee scrutinises proposals for new regulation (“regulatory provisions”) by combing through the analysis and assumptions that civil servants make before the proposal becomes legislation.

The RPC isn’t evaluating whether the policy is a good or bad policy – that’s a political question for Parliament to decide. Instead, it looks at how well-evidenced the costs and benefits used in the policy are and points out areas where the analysis needs to be stronger. The Committee doesn’t get a veto on any policy, but it does publish its evaluation of the Options Assessments and Impact Assessments which government departments write.

The quality of the analysis in these Assessments matters, because once regulation is created it is rarely revised. There’s a kind of “regulatory ratchet” in Britain that means lots more is created than is ever removed. The Government’s target of reducing the administrative burden from regulation is part of the answer to that. But so is getting regulation right the first time, and that’s where the RPC comes in.

It matters even more because the costs and the benefits of regulation rarely fall on the Government itself – they mostly affect businesses and the public, and often in ways that are hard to trace back to the change in regulation.

To give you a sense of scale, small changes from regulation can have a big impact when they are applied across the whole economy. In line with the Treasury’s Green Book, the costs and benefits of regulation are evaluated like other policies: all the costs and benefits are both quantified with a financial value (including those which aren’t an actual financial cost or benefit, but are still given an approximate value), and measured against each other to weigh up whether the regulation is a good idea. Measures like Net Present Social Value capture all the costs and benefits, including indirect ones, and help us understand if regulation is worth it – because often the costs are much, much greater than the benefits.

Take the Second Staircase rule, brought in following the Grenfell Tower fire, which cost over £2.6 billion more than the benefits it brings according to the civil service’s own Impact Assessment. Or the Terrorism (Protection of Premises) Act, also known as “Martyn’s Law” (brought in following the Manchester Arena bombing), which will have a negative Net Present Social Vale of £1.8 billion to the economy. Both are measures designed to protect the public, and incorporate the impact of saving lives in the benefits case, but these are still significantly outweighed by large costs to everyone else. When the costs from regulation can be on that scale, even greater in many cases, good evidence is essential for making the right call.

Often, the analysis government departments produce just isn’t good enough. The National Audit Office’s review in 2012 of attempts to reform the Tier 4 student visa route found that the Home Office had significantly underestimated the costs to education providers, by omitting whole areas of some costs and getting others completely wrong. One-off familiarisation costs, which the Home Office had thought would be about £25, were in fact over £500 per provider. Mandatory costs from new oversight measures were entirely excluded, amounting to between £9,000 and £20,000 per organisation in the first year alone. Whole areas of cost, like verifying English language requirements, were also omitted altogether. The NAO calculated that the regulatory burden would be over £40 million a year, far higher than the Home Office’s estimated £25 million.

The RPC checks the Government’s analysis

These are the kind of problems the RPC look for in the Options Assessments and Impact Assessments. The Better Regulation Framework requires departments like the Home Office to submit an Options Assessment to the RPC for an opinion before the minister decides on a preferred option, so the RPC can advise on the methodology used to decide between different approaches to the problem.

The Committee classifies these assessments in two ways:

• Green (fit-for-purpose): The RPC has no significant concerns over the quality of the assessment, or there are some minor issues that could be improved. There may be many points for improvement, which the department should consider.

• Red (not fit-for-purpose): The RPC has major concerns over the quality of the evidence and analysis, and the overall quality of the assessment that need to be addressed.

It gives a green or red rating to three different areas of an OA: the rationale, the identification of options, and the justification of the preferred option. If any of the three is rated red, then the Assessment is treated as not fit for purpose. Alongside these areas, it assesses the quality of two more sections – the ‘Regulatory scorecard’ (qualitative and quantitative assessments of the impact of regulation) and the Monitoring and Evaluation plans. These areas are rated using a four-point scale of “good”, “satisfactory,” “weak”, or “very weak.” They don’t affect the formal Green/Red rating but highlight where analysis is good or could be improved.

Where an OA is likely to receive a Red rating, the RPC can issue an Initial Review Notice (IRN) to the relevant department requesting further work prior to issuing a formal opinion. In the 2024-25 year six proposals (22 per cent of the total) – received IRNs. All six updated their submissions and received ‘fit-for-purpose’ ratings.

The Government’s Options Assessments themselves aren’t routinely published, although some are when they are used as part of a public consultation on which option to pursue – for example, the Make Work Pay: duty to inform workers of their right to join a union which was published last summer. Along with different options for how to deliver the policy intent, all OAs include a “do nothing” option, which compares the costs and benefits of not acting with the preferred option.

Once an option has been picked, the Government conducts a full Impact Assessment, taking on board the RPC’s feedback on the OA. These IAs are generally published at the same time the legislation is laid in Parliament. The RPC used to scrutinise all IAs instead of OAs, but now it scrutinises OAs it only looks at some IAs. It evaluates larger measures (with estimated costs of greater than £100 million) where either the RPC’s scrutiny of the earlier Options Assessment showed it was weak, or where it falls in a policy area which is exempt from scrutiny at OA stage.

Whole areas of regulation are entirely exempt from scrutiny at OA stage, such as measures to ensure Government complies with international obligations or court judgements. And all regulatory provisions related to issues of building safety are entirely exempt from all elements of RPC scrutiny – a decision taken following the Grenfell fire to smooth the passage of building safety regulations.

The point of the RPC publishing their analysis of OAs and IAs is that it makes the Government take their feedback seriously – in the same way that the Office for Budget Responsibility (another important quango) publishes an independent economic analysis which the Treasury uses at a Budget.

And the Government would do well to take that feedback more seriously, because there are plenty of areas for improvement. The proportion of assessments rated ‘good’ or ‘satisfactory’ for their rationale section fell from 84 per cent in 2022 to 71 per cent in 2024. The cost-benefit analysis section’s ‘good’ and ‘satisfactory’ ratings declined from around 90 per cent to 76 per cent over the same period; and satisfactory assessment of wider impacts fell from roughly three-quarters of submissions to below 60 per cent.

In other words, government analysis of regulation is getting worse. And it is even worse in some areas than others. 60 per cent of the Treasury’s Impact Assessments were deemed unacceptable by the RPC on first review and required substantial revision before they can be considered – along with 44 per cent of DEFRA and DHSC assessments.

When assessments are not deemed fit for purpose, that means the RPC thinks there’s a sufficiently big area for improvement. When the RPC initially reviewed the IA for the regulatory implications of the UK-India Free Trade Agreement in summer 2025, it was initially rejected. The RPC found problems with the modelled impacts on both UK GDP and greenhouse gas emissions, but these were corrected by the final IA so the RPC could rate it fit for purpose.

Where the RPC believe the Government’s assessment is not fit-for purpose, and it received a Red rating, then according to the Better Regulation Framework the Government can either resubmit their assessment with changes or go ahead anyway while making clear the RPC’s concerns. The RPC track the small number of instances departments have chosen to proceed with a regulatory provision with the RPC finding the assessment is not fit for purpose. Martyn’s Law (mentioned above) is one example, where the RPC said the IA did not provide evidence that the proposal would reduce the risk of terrorism at smaller venues, and didn’t sufficiently quantify the economic impact of regulation on those smaller venues.

The RPC also evaluates Post-Implementation Reviews (PIRs), which the Small Business, Enterprise and Employment Act 2015 requires departments to do within five years of implementing any new regulation. Last year the RPC also started tracking whether government departments are doing the PIRs quickly enough to comply with the Act, and publishes the size of the backlog – as of December there were 47 outstanding PIRs, which had come down from a total of 80 when the RPC first started reporting in April.

The RPC shows how quangos can work well

The RPC is a good example of a quango set up for the right reasons – to use its day-to-day independence from ministers to drive a Government priority (in this case, better regulation) through correcting for a structural imbalance: although most governments say they want better regulation, in practice the incentives to create regulation regardless of its quality are very strong. The system needs a way to correct for that.

Lots of quangos aren’t playing this kind of useful role. When we wrote Quangocracy last year, we assessed that over 100 public bodies should not be independent, and should either go back under the direct control of ministers, be scrapped altogether or be reconstituted in some way. But the RPC isn’t one of them – because its independence is a crucial part of establishing the integrity of figures used by the Government. Like the OBR, or the Office for National Statistics, having independent scrutiny of public data is crucial to good policymaking and democratic accountability. A Government which didn’t have public scrutiny of regulatory cost-benefit analysis would find it impossible to control the quality of the regulation it created.

Better regulating for better regulation

The RPC’s role should be strengthened. There isn’t a statutory basis for its work, or what Government departments must share with it – just the Better Regulation Framework guidance. A legislative footing for its responsibility to assess the evidence behind regulatory provisions, and publish the results, would make it much harder for government departments to ignore – like the OBR’s role with the Treasury, or the Social Security Advisory Committee’s role with the Department for Work and Pensions.

It doesn’t make sense that some sections of the RPC’s analysis can get a department a Red rated Impact Assessment if they do a bad job, but not all of them. The monitoring and evaluation section is one of the more frequently ‘weak’ or ‘very weak’ rated sections, but that isn’t sufficient to get an Impact Assessment rated Red. All sections should carry the same weight, and make the Government turbo-charge their improvement plans.

And it’s an anomaly that the RPC scrutinises some kinds of regulation and not others. Areas like building safety regulation should be treated the same as every other area. And they should scrutinise all Impact Assessments even if they have already looked at an Options Assessment for the same policy (and publish the results).

This kind of independent scrutiny does add friction into policymaking – it slows things down, and constraints the Government’s ability to act. In lots of areas of public policy we’ve made it far too hard for governments to act decisively – we’ve written about how the State needs much less friction in hiring and firing, technology adoption, budgeting and procurement.

But in regulation, we have the opposite problem. It has become far too easy to create new regulation, so governments default to it in place of better policies. Indeed, perhaps regulation has been over-used precisely because we’ve made it so hard for governments to act decisively in other ways. Politicians can quickly decide to regulate – too often with minimal benefit and significant (but hidden) costs – just to show that they are ‘doing something’.

Along with making it easier for Government to act decisively in other ways, it’s important to raise the bar on the quality of analysis that regulation needs.

In Thinking the Twentieth Century, the historian Tony Judt wrote that:

“From such a distance one can make arguments about history’s costs and benefits: but the costs are borne by someone else and the benefits can be anything you wish to imagine.”

He could easily be talking about policymakers and the future. The benefits of bad regulation can be anything politicians wish to imagine, and the costs are always borne by someone else.

Strengthening the independence of the RPC is a crucial part of correcting that.