UK Statistical System UK Statistics Authority Code of Practice
UK Statistics Authority Correspondence

Penny Young to Julie James MS: Welsh apprenticeships

This letter was sent from Penny Young, Interim Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Julie James MS 

Thank you for your letter of 26 March, which responds to my previous letter, on the use of statistics to monitor progress against the Welsh Government’s target of creating 100,000 apprenticeships in this Senedd term. 

In this case, the relevant official statistics and management information on apprenticeships are produced by Medr. Your letter explains that recent public statements relating to apprenticeships have drawn on management information, rather than official statistics, because management information provides the most up to date picture.  

As you are aware, Medr publishes two measures of apprenticeship starts, based on management information. The first, more stringent, measure excludes apprentices who withdraw in the first 8 weeks, and those who transfer out of their programme to other learning. In Medr’s published statistics this is referred to as the rigorous measure. The second measure records all apprentices who have started an apprenticeship. Further to your letter we acknowledge that both measures go through the same quality assurance measures, and both will have appropriate uses. 

Your letter explains that recent public statements have drawn on the measure that records all apprentices who have started an apprenticeship, and the reasons for this. It is helpful to have this explanation, which was not given alongside recent public statements, especially where there is more than one possible measure available.

Based on public information available about monitoring the target, use of the measure that records all apprentices who have started an apprenticeship appeared to create an inconsistency with previous reporting.

In the Quality and Methodology section the official statistics state that: 

The statistics for the target measures use a more rigorous measure of apprenticeship programme starts than other statistics in this output. This measure takes account of early drop outs (within first 8 weeks) and transfers between apprenticeships.

The StatsWales publication Data on apprenticeships started in relation to target measure highlights the same point. 

The most recent management information at the time of my earlier letter, published on 26 February, set out the position in relation to the Welsh Government’s target: 

The total number of apprenticeships started in the Senedd period crossed 100,000 in January 2026. The number of apprenticeship starts for the more rigorous target measure remained below 100,000. 

It was on this basis that I wrote to the First Minister earlier this month.

Going forward, I would strongly encourage ministers to be transparent about which measure they are using and why, considering any previous precedents that might influence the understanding of those receiving the statement.

I hope this letter more fully explains our position, and I welcome your commitment to the appropriate and transparent use of statistics. In line with that, I would always encourage Ministers to seek the advice of their statisticians on public statements related to official statistics. 

Yours sincerely,

Penny Young
Interim Chair

Letter from Penny Young to Rt Hon Eluned Morgan MS – Welsh apprenticeships

Dear Eluned Morgan MS, 

As Interim Chair of the UK Statistics Authority, which oversees the Office for Statistics Regulation, I am writing to you regarding the use of apprenticeships statistics in a press release on 9 February 2026 titled “Over 100,000 apprenticeships delivered in Wales this Senedd term” and in the Senedd during First Ministers questions on 24 February 2026. 

In both cases the claim was made that Welsh Labour has achieved its target of having delivered 100,000 apprenticeships in the last four years.  

There are two measures of apprenticeship starts published by Medr, Wales’s Commission for Tertiary Education and Research. The well established ‘rigorous’ measure excludes apprentices who withdraw in the first 8 weeks, and those who transfer out of their programme to other learning. The alternative measure records all apprentices who have started an apprenticeship. 

According to Medr’s management information, as at 26 February 2026 there were 92,800 apprenticeship starts since Q4 2020/21 using the rigorous measure of progress towards Welsh Government’s target. When including the starts which are not counted as part of the more rigorous target measure, there were 101,760 apprenticeship starts overall in the time period.

The rigorous measure has been used historically to report on progress, including in annual reporting during the current Senedd term. This was also the measure used prior to the target being reduced from 125,000 to 100,000 in June 2024. 

The claim that over 100,000 apprenticeships have been delivered in Wales in this Senedd term is not supported by the rigorous measure and appears to create an inconsistency with previous reporting and could be perceived as cherry-picking the data. The press release and statement during First Minister’s questions were not clear that they were based on the alternative measure that includes all apprentices. 

Transparency matters because it enables debate to focus on the important issues. It is crucial to avoid using data that is overly selective or missing appropriate context.

Being clear on definitions and the evidence underpinning any such statements can help maintain public confidence in statistics and the organisations that produce them. 

Yours sincerely,

Penny Young
Interim Chair

Penny Young to Sir James Cleverly MP, Andrew Griffith MP and James Wild MP – changes to Business Rates

This letter was sent from Penny Young, Interim Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Sir James, Mr Griffith, and Mr Wild,

Thank you for your letter of 8 January 2026 about the changes to business rates announced in the Budget of 26 November 2025.

Your letter focused on the claim that the “tax rates” for business rates for Retail, Hospitality and Leisure (RHL) business properties are at their lowest since 1991. You cite several sources where this is quoted and argue that the claim is misleading for several reasons, which we evaluate in this letter.

First, you highlight that there are two components to business rates: the multiplier and the Rateable Value of the property. Comments about the tax rates in the documents and statements you shared only reference the former and not the two taken together; and the increase in rateable value means an increase in the amount of tax due for many businesses. Our assessment is that it is correct to assert that while the new multiplier rates are lower (and in some cases the lowest since 1991) in 2026/27 compared to 2025/26, business rates overall are likely to rise for many RHL businesses because of Rateable Values owing to the revaluations of 2024 due to come in from April 2026. Our conclusion is that to support public understanding, statements should have been clearer as to whether they referred to changes to the tax rates (i.e. multiplier) specifically or the changes to business rates more broadly, as the tax paid as a result of the latter might increase despite the reduction of the multiplier.

Second, you argue that the comparison to 1991 only applies to one of the new business rates announced in the Budget speech. We have confirmed that it is only the new RHL multiplier for small businesses that sees its lowest tax rates since 1991. While the other RHL multipliers (including the new rate for higher value premises above £500,000) will fall from 2025-26 to 2026-27, it is not to the lowest level since 1991 for these other RHL categories. The Chancellor of the Exchequer’s speech you cited also referenced

“lower tax rates for over 750,000 retail, hospitality and leisure properties—the lowest rates since 1991”.

It is our understanding that the 750,000 figure refers to all RHL properties, who will see tax rates fall but not necessarily to the lowest levels since 1991.

Thirdly, you argue that it was necessary to factor in other changes made in the policy area, in particular (but not limited to) the reduction and subsequently planned abolition of the Retail, Hospitality and Leisure (RHL) Rate Relief that had been introduced by the previous Conservative Government. It is true that these changes may increase the tax paid by businesses.

Fourthly, you mention that a new surcharge is being applied to higher value premises with Rateable Values above £500,000, which you argue is hitting twice as many retail premises than the online warehouses this was supposedly meant for. We agree it would have been beneficial for involved Departments to present this fuller context proactively among their suite of post-Budget information to support understanding.

Finally, you cite the reference to tax rates made in the HM Treasury Retail, Hospitality and Leisure Factsheet. While the factsheet mentions a tax cut, our assessment is that the factsheet does a good job of setting out the changes to business rates in their fullest context. This includes an acknowledgement that

“this fall in the tax rate isn’t big enough to offset the impact of new post-COVID valuations, plus the ending of the temporary relief that many pubs and other RHL businesses benefit from that has been winding down since COVID.”

It also uses worked examples to illustrate that, while some of its “big protections” like the “Transitional Relief” you cite cap losses, they do not eliminate them. Overall, it is our view that this is an example of good statistical communication.

Taking these examples together, we agree that there were opportunities for improvements to be made to support understanding of the data and avoid the potential for people to be misled.

It is welcome that the Treasury factsheet has provided a good level of detail on the impact of all changes to business rates. It would have better supported public understanding if ministers had been just as clear that although they have cut tax rates, they have also introduced substantial transitional measures to support businesses who would otherwise be liable to pay, in the Government’s own words, “big increases” in tax as a result of the abolition of the RHL relief.

We have passed on our findings to HM Treasury and the Ministry of Housing, Communities and Local Government to ensure our expectations for transparent communication are met in future announcements involving statistics.

Thank you again for raising this matter with us.

Yours sincerely,

Penny Young
Deputy Chair

 

Penny Young to Kevin Stewart MSP: Waiting times in England and Scotland

This letter was sent from Penny Young, Interim Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Mr Stewart,

Thank you for your letter of 20 November regarding a social media post and video shared on X.com by Michael Shanks MP about waiting lists in Scotland.

The UK Statistics Authority, through its regulatory arm the Office for Statistics Regulation (OSR), has investigated this issue and I have today written to Mr Shanks to share OSR’s conclusions and our broader expectations about the use of statistics in public debate.

I would like to thank you for bringing this issue to our attention.

Yours sincerely,

Penny Young
Deputy Chair

Penny Young to Michael Shanks MP: Waiting times in England and Scotland

This letter was sent from Penny Young, Interim Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Mr Shanks,

I am writing to you regarding a post and video you shared on X.com about waiting lists in Scotland, following concerns raised with the Office for Statistics Regulation (OSR), the regulatory arm of the UK Statistics Authority (the Authority).

OSR investigated the following three claims:

  1. That waiting lists have been altered (via “clock resets” – where a patient’s waiting times clock is reset to zero during their wait) to make the numbers look better than they are.
  2. That “if you’ve been waiting for an appointment, for some reason, the first date you’re given, you’re not able to make, well the clock gets reset”, and that this reduces waiting lists even though those patients are still awaiting treatment.
  3. That over 12,000 people were waiting over two years for an NHS appointment in Scotland, and that in England, there were 168.

In 2023 the Scottish Government introduced new guidance on calculating waiting times. The guidance included new rules about “clock resets” and how they impact the calculation of waiting times. Public Health Scotland (PHS) produces monthly accredited official statistics on stage of treatment waiting times for inpatients, day cases and new outpatients. After two years of work with Scottish health boards to implement the new guidance, PHS updated how it calculates waiting times in October 2025. In its recent official statistics, and an accompanying blog, PHS explained that under the new rules, waiting times estimates are marginally shorter than was previously the case.

In relation to the first claim about alterations to waiting lists, it is not for the Authority to comment on Scottish Government policy decisions such as the changes to waiting times guidance. However, we are satisfied with the way that PHS has implemented this. We consider that the change to waiting times calculations in its official statistics complies with the Code of Practice for Statistics, is appropriate, and supports user understanding. PHS wrote to OSR in August, notifying it of the upcoming change, and OSR responded supporting this decision.

In relation to the second claim, the Scottish Government’s 2023 guidance states that the clock can only be reset if a patient does not attend an agreed appointment, or following the refusal of two or more reasonable offers of appointments. These clock resets do not impact total waiting list sizes, as explained by PHS in its impact assessment, which says that “waiting list sizes (ongoing waits) are unaffected by the revised guidance”.

In relation to the third claim, sources for the numbers were not cited together with the post on X.com as we would normally expect under the Standards for the Public Use of Statistics, Data and Wider Analysis. You later confirmed to us that the numbers come from PHS’s official statistics on waiting times and NHS England’s official statistics on referral to treatment waiting times. The total of people waiting over two years for treatment in Scotland was created by adding together the number of ongoing waits longer than 104 weeks for a new outpatient appointment, and the number of ongoing waits longer than 104 weeks for an inpatient or day case admission.

These two numbers should not be added together. This is because the same person could be waiting for both an outpatient appointment and an inpatient or day case admission. As PHS explains in its publication, “to avoid overestimating…the number of ongoing waits should not be added together”.

It is also important to note that waiting lists in both Scotland and England, are reported for treatment ‘pathways’, not patients. This means that the figures could include the same person more than once if they are waiting for more than one treatment and do not necessarily represent the total number of people waiting. This issue was further explored by PHS in a blog last month.

We know that many users want to be able to compare NHS performance across the UK and that it is frustrating for them when they are unable to do so. We therefore support and encourage efforts to produce more comparable data where possible. In this case, however, producers have explained that for planned care, different healthcare policies, commissioning processes and patient data systems make it particularly challenging.

The ONS published work on the UK comparability of waiting times statistics in July 2024 that highlights differences in data collections and definitions across the UK. These include differing policies between Scotland and England, for example on consultant vs non-consultant led pathways, clock stops and adjustments and transfers to and from private settings. Due to these differences, the ONS and producers of health statistics across the UK are clear that their data cannot be used to make direct comparisons of waiting lists between nations in this way. ONS does advise that it can be helpful to compare trends in waiting times statistics between England and Wales, and separately between Scotland and Northern Ireland, because the systems and compilation of statistics in these nations, while still different, is most comparable. OSR has asked PHS to make this clearer in its regular publications.

Waiting times are of high public interest across the UK and it is vital that statistics about them are used appropriately to support public debate. We accept that the public and users will be frustrated about the limitations in comparability. But to uphold public confidence in statistics, we encourage you to ensure statements containing statistics are presented clearly and with sufficient context to avoid the potential for people to be misled.

Yours sincerely,

Penny Young
Deputy Chair

Professor Dame Carol Propper to Seamus Logan MP: Statements on water quality

This letter was sent from Professor Dame Carol Propper, and is also available on the UK Statistics Authority website.

Dear Mr Logan,

Thank you for your email of 5 September to the Acting National Statistician regarding comments by the former Secretary of State for Environment, Food and Rural Affairs, that drew comparisons between England and Scotland’s water industry regulation.

I respond on behalf of the UK Statistics Authority’s Regulation Committee, which oversees the independent regulator, the Office for Statistics Regulation (OSR). Whereas normally the Chair of the Authority would respond to your concern, she has instead asked me to respond in her place, to avoid the risk of a perceived conflict of interest with another role.

Judgements about regulation of the water industry and broader environmental policy are rightly for you and your colleagues in the UK Parliament and devolved assemblies to consider. However, we have looked at some of the statements made as part of our role enforcing the Code of Practice for Statistics; a full list of these is enclosed in Annex A.

In brief, we consider that these statements lacked enough transparency about their sources to be verified, and that the broad evidence did not support them. Without appropriate discussion of the limitations of some of the more specific figures quoted, they run the risk of misleading the public.

As you point out, the then-Secretary of State said several times that water pollution and water quality are worse in Scotland than in England. He made these claims repeatedly in the House of Commons, as well as in broadcast interviews and online. We also note his letter of 24 July to Stephen Flynn MP published on X, which cites a range of statistics in response to the dispute over his claims.

The table of figures appended to the letter is not relevant to water pollution and water quality, but in the body of his letter he further claimed, without providing a source, that in 2023-24, there were 35.8 pollution incidents per 10,000km of sewerage network in Scotland, and 35.4 in England and Wales combined.

The Department for Environment, Food and Rural Affairs told us that its source for Scotland was an analysis of Scottish Water’s annual report, and for England and Wales, Ofwat’s water company performance report, wherein pollution incident rates are broken down by company. The sector average given for both England and Wales combined is 35.42. These water companies’ territories do not neatly match national boundaries, so it is not simple to calculate pollution incidents in England alone.

Underlying regulatory data on pollution incidents are not yet available for Wales or Scotland, although an Environmental Standards Scotland report claims Scottish Water has a lower incident rate (36) than English water companies (41), when including incidents from water supply assets. However, there are many differences in monitoring and reporting of pollution incidents data, so the Independent Water Commission (IWC) cautions against making these comparisons between countries. It is important to provide this context when referring to them in public debate.

The IWC concluded its review of the water sector on 21 July, and in its final report there are several sets of statistics directly relevant to water quality and other environmental data. It would have greatly improved public understanding to draw upon these figures, given their relevance, quality, and importance to the Government’s stated ambitions. The figures show broadly that Scotland has a similar or better share of bathing and surface water sites that meet good or excellent standards for water quality. The OSR has published a further analysis of this topic, which I enclose in Annex B.

The Authority expects that ministers take care to avoid using data that is overly selective or missing appropriate context. Based on the statements made without discussion of their context, sources, and limitations, there is the potential for people to be misled about English and Scottish water quality and infrastructure. As former Chairs have said, omitting this kind of information can damage public trust in the data, so we encourage those speaking on behalf of Government to ensure statistical statements are presented clearly and transparently, in a way that supports public understanding.

 

Yours sincerely,

Professor Dame Carol Propper
Chair of the Regulation Committee

ANNEX A

  • “[…] you were just hearing about Scottish Water, they’re nationalised, pollution in rivers in Scotland is worse than in England”
    Sunday with Laura Kuenssberg, BBC One, 20 July
  • “In Scotland they have a nationalised water company, but pollution levels in Scotland are worse than they are in England“
    Channel 4 News, 21 July
  • “Scotland has a nationalised water company… and water pollution is worse than in England as a result”
    House of Commons, 21 July and republished online at X.com
  • “… under the nationalised model in Scotland, pollution is worse, not better.”
    House of Commons, 21 July
  • “Official statistics […] show several areas where Scotland’s water quality underperforms relative to England.”
    Letter to Stephen Flynn MP, 24 July
  • “There were 35.8 incidents per 10,000km of sewer in Scotland versus 35.4 incidents per 10,000km of sewer in England and Wales reported in 2023-24.”
    Letter to Stephen Flynn MP, 24 July
  • “OFFICIAL: The SNP are managing water pollution in Scotland even worse than the Tories did in England”
    @SteveReedMP on X.com, 25 July
  • “Levels of pollution in England are bad enough, but under the SNP in Scotland they are even worse”
    House of Commons, 4 September
  • “I have published the data and I stand by it: pollution under the SNP in Scotland is even worse than it was under the Tories in England.”
    House of Commons, 4 September

Annex B

OSR statement on statistics concerning aspects of water in Scotland and England

Letter from Sir Robert Chote to Jeremy Balfour MSP – Health and Wellbeing Census Scotland

This letter was sent from Sir Robert Chote, Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Mr Balfour

Thank you for your letter raising your concerns about the Scottish Health and Wellbeing Census (HWBC).

Many of the issues about data governance and sharing procedures discussed by your constituents are out of remit for the Office for Statistics Regulation (OSR) and fall within the remit of the Information Commissioner’s Office (ICO). However, I have addressed your comments relating to the OSR’s letter to the Scottish Government, published in July 2022.

This letter set out that the Scottish Government needed to review the approach it took to question development including the “legal and ethical governance arrangements that are in place for asking questions of each age range”. Please note that the OSR did not request the Scottish Government to undertake a review of the survey and data governance processes as a whole.

It is regrettable that the outcomes of this review have not yet been made public and the OSR has now written to the Scottish Government to set the expectation that this is rectified within the next 30 days.

Yours sincerely,

Sir Robert Chote
Chair

 

Related links

Ed Humpherson to Alastair McAlpine: Scottish Health and Wellbeing Census

Jeremy Balfour MSP to Ed Humpherson – Health and Wellbeing Census Scotland

Letter from Sir Robert Chote to Paul O’Kane MSP: Scottish Government child poverty statistics

This letter was sent from Sir Robert Chote, Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Mr O’Kane,

Thank you for your letter expressing concern about statements made by both the current and previous Scottish First Minister that Scottish National Party (SNP) policies are “lifting” an estimated 100,000 children out of poverty.

Under the principles of intelligent transparency, it is always advisable to think how an average person would interpret a high-profile quantitative claim of this sort so as to minimise the danger of them being misled and thereby trust in similar statements being undermined.

In this instance, the First Ministers were referring to a modelled estimate of the difference between the level of relative child poverty expected in 2024/25 and the level that we would have seen in the absence of a number of policy measures (as set out in footnote 4 of the updated Cumulative Impact Assessment). The Scottish Government set out the methodology behind this comparison in an annex to its Cumulative Impact Assessment in 2022 and the 100,000 estimate comes from the update published in 2024.

This kind of analysis is a reasonable way to estimate the impact of Scottish Government policies on child poverty, even though, just like alternative estimates, the calculations are bound to be uncertain and dependent to some degree on methodological choices. But the average person hearing such a statement might well assume that the First Ministers were claiming that child poverty is 100,000 lower than when the SNP took office. And, as you point out, the Scottish Government’s official statistics on Poverty and Income Inequality in Scotland conclude that the proportion of children in Scotland living in relative and absolute poverty remains broadly stable. Comparing the number of children living in relative poverty from 2004-07 (pre-SNP) to the latest datapoint of 2020-23 shows a decrease of 10,000 children (250,000 to 240,000). For children living in absolute poverty the decrease is 40,000 children (250,000 to 210,000).

Given this potential confusion, Ministers would be well advised from time to time to accompany this type of claim with a reminder of the methodology underpinning it so that they are not suspected of making an unduly flattering comparison.

Yours sincerely,
Sir Robert Chote
Chair

Related links:

Paul O’Kane MSP to Sir Robert Chote – Scottish Government child poverty statistics

 

Sir Robert Chote to Rt Hon Richard Holden: Party spending claims

This letter was sent from Sir Robert Chote, Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Mr Holden,

Thank you for your letter of 6 June regarding the Labour Party’s analysis of Conservative Party commitments and its own plans.

In my recent letter to political parties, I asked that parties and candidates use statistics appropriately and transparently during this general election campaign and made clear why these expectations are in the interests of the public and those campaigning. These expectations were echoed in a statement published by the Office for Statistics Regulation regarding the claim that “a Labour government would mean £2,000 of tax rises per working household”. Many of the principles set out in that statement apply also to the claim you highlight, that the Conservatives have “£71 billion of unfunded spending plans”.

This figure derives from Labour’s 25 May document Conservatives’ Interest Rate Rise which sets out their costings of nine future ‘policy decisions’ and refers to roughly £71 billion of net extra spending in fiscal year 2029-30. In another document, Tory Manifesto Costings published on 13 June, the Labour Party claimed that Conservative manifesto plans would amount to net extra spending of roughly £71 billion over the next five fiscal years put together and “raise people’s mortgages by £4,800” cumulatively over that period.

Future costings are always subject to uncertainty and dependent on choice of methodology. To help people understand the assumptions that have gone into costing models, it is essential that the underlying calculations, data sources and context are provided alongside the figures. When distilling these claims into a single number, there should be enough context to allow the average person to understand what it means and how significant it is. Omitting this information can damage trust in the data and the claims that these data inform.

To safeguard trust in official statistics, we encourage that statistical claims are presented clearly and transparently so that the public can test the arguments, and descriptive statements, that political candidates make about them.

Yours sincerely,
Sir Robert Chote

Letter from Sir Robert Chote to Lord Bailey of Paddington AM: Housing statistics

This letter was sent from Sir Robert Chote, Chair of the UK Statistics Authority, and is also available on the Authority website.

Dear Lord Bailey,

Thank you for your letter to the National Statistician regarding the Mayor of London’s use of housing statistics in a LabourList article from March 2024. This was passed to the Office for Statistics Regulation (OSR), the regulatory arm of the UK Statistics Authority.

The Mayor’s comments appear to draw from the Greater London Authority’s (GLA) Affordable Housing starts and completions statistics. The statistics are broken down into house building ‘starts’ and ‘completions’. These definitions are clear and describe a real-world picture of what the statistics represent: whether a dwelling has started construction or whether construction has finished.

There were 25,658 affordable housing starts in the period April 2022 to March 2023, and 13,954 affordable housing completions during this time. The Mayor’s use of the word ‘delivered’ could be misinterpreted by the average person to mean the housing had been completed rather than started. While the Mayor would not have had access to the statistics for the equivalent period in 2023/24 at the time of writing the article, the number of affordable housing starts and completions for 2023/24 was 2,358 and 10,949 respectively. The reference to ‘within the last year’ may have been intended to reflect the latest year of statistics that were available, but this would not be clear from the statement alone.

In line with the principles of intelligent transparency, when making numerical claims, public figures should be clear what they are referring to and consider how a reasonable person would interpret the claim. To help the public understand statistical claims made during the General Election debates, the OSR has recently published a series of explainer articles including on housing supply and affordability statistics.

Yours sincerely,
Sir Robert Chote
Chair

 

Related links

Lord Bailey of Paddington AM to Sir Ian Diamond – housing statistics