Ed Humpherson to Steve Ellerd-Elliot: Assessment of the Department for Work and Pensions’ Benefit Statistics

Dear Steve

Assessment of the Department for Work and Pensions’ Benefit Statistics

We have today published the assessment report on these statistics. I am grateful for the positive engagement of your team throughout the assessment process.

The DWP benefit statistics publication brings together key statistics on DWP administered benefits and acts as a compendium of benefit statistics, where the frequency and reference date vary across the benefits. The statistics provide valuable insight into the changing welfare system. There has been heightened interest during the COVID-19 pandemic and DWP has shown its ability to respond and adapt to changing demands by providing real-time insights into the take up of Universal Credit.

We found that DWP has a range of policies in place that ensure strong data governance, including an information management policy and a policy statement on confidentiality and access. These measures enhance trustworthiness and give users confidence that the data are being managed securely.

The report also highlights some improvements that we consider necessary to enhance the public value of the statistics. Making these improvements is necessary to ensure that these statistics can continue to be badged as National Statistics.

Our findings point to a lot of untapped potential in the statistics. We identified a strong need for DWP to engage more proactively with users to understand their needs, rather than relying on users reaching out to the team. Insufficient user engagement has led to developments not being aligned with priorities for users and key data gaps remaining unfilled, embedding more-strategic user engagement into the production and development of the statistics would ensure they keep pace with statistics on Universal Credit and remain valuable to users.

Information about the quality of the statistics does not support users’ understanding of the strengths and limitations of the statistics and risks the data being misinterpreted. Signposting between Stat-Xplore (DWP’s interactive tabulation tool) and the DWP landing pages for the statistics needs to be improved to ensure users are equipped to interpret and use the statistics appropriately.

Yours sincerely

 

Ed Humpherson

Director General for Regulation

 

 

Ed Humpherson to Steve Ellerd-Elliott: Confirmation of National Statistics Designation for National Insurance Number Allocations to Adult Overseas Nationals Entering the UK Statistics

Dear Steve,

Confirmation of National Statistics Designation for National Insurance Number Allocations to Adult Overseas Nationals Entering the UK Statistics

 In January 2017 the Office for Statistics Regulation undertook an assessment of Statistics on National Insurance Number Allocations to Adult Overseas Nationals produced by the Department for Work and Pensions (DWP). We assessed the statistics reported quarterly in National Insurance Number allocations to adult overseas nationals (NINo Allocations), and annually in Nationality at point of NINo registration of DWP working age benefit recipients (NINo Benefit Claimants), along with the related commentary that is published in the August publication of NINo Allocations.

As a result of the assessment, we judged that the statistics published in NINo Allocations should have their National Statistics status temporarily suspended until work to improve their limited supporting guidance and overall public value is undertaken. We have reviewed the actions that your team has taken since then to address the requirements in the assessment report. On behalf of the Board of the UK Statistics Authority, I am pleased to confirm the designation of National Statistics for the NINo Allocation statistics.

The 2017 assessment of the statistics focused on eight requirements which DWP was to meet to ensure compliance with the Code of Practice for Statistics.  We also considered more broadly how the statistics meet the three pillars of the Code of Practice for Statistics, evaluating their trustworthiness, quality, and value.

We recognise the hard work your team has undertaken to meet the requirements of the assessment which has enhanced the clarity, insight, and accessibility of the statistics for users. From our review, users have told us that improvements made to the statistics have strengthened perceptions of trustworthiness. Further, application of the Quality Assurance of Administrative Data (QAAD) guidance in the related supporting information provides detailed information and assurance of the quality of the statistics.

At the time of the assessment we stressed the importance of the role of DWP and these statistics in helping to inform what we understand of the wider migration picture in the UK. This will be even more important in this period of change to the migration system in the UK and with the cross-government migration statistics transformation programme led by ONS. We welcome DWP’s commitment and efforts to support ONS in taking forward the transformation programme.

The review focused solely on NINo Allocations; NINo Benefit Claimants is still classed as experimental statistics so is not being considered for National Statistics designation at this time. I am encouraged to hear that, working in collaboration with HM Revenue and Customs (HMRC), you plan to develop these statistics to include both DWP and HMRC benefits. This should provide users with a more complete picture of working age benefit recipients by nationality.

We have included more detail about our judgement in an annex to this letter, which I, or a member of my team, would be very happy to discuss with you in more detail.

I have copied this letter to Yolanda Ruizrodriguez in your team.

Yours sincerely

 

Ed Humpherson

Director General for Regulation

 

 

Annex: Review of actions taken by Department for Work and Pensions in response to Assessment Report 331

Please note in the Annex, the National Insurance Number Allocations to Adult Overseas Nationals Entering the UK statistics will be referred to using the acronym ‘NINo’.

Annex: Review of actions taken by Department for Work and Pensions in response to Assessment Report 331  
Please note in the Annex, the National Insurance Number Allocations to Adult Overseas Nationals Entering the UK statistics will be referred to using the acronym ‘NINo’.
RequirementActions taken by DWP to meet requirementsOffice for Statistics Regulation’s evaluation of evidence in meeting requirements
Value
Requirement 1:
To maximise the public value of the statistics DWP should improve the commentary in the NINo statistical reports by:
a) Adding more context, including explaining how they feed into wider analysis on immigration data
b) Providing more explanation of changes seen in the time series, in particular the reasons for the large increases in NINo registrations since 2014
c) Improving the signposting of the commentary for NINo Benefit Claimants or merging NINo Allocations and NINo Benefit Claimants into one publication
a) In February 2020, as outlined in DWP’s release strategy, additional context was added to the quarterly report to better explain the historic trends in data. DWP also published a background information note, which details the purpose of the statistics, their limitations and how they fit in with other sources of migration data.
b) DWP has added an explanation of changes seen in the time series to the bulletin, including warning users of the changes to the process of NINo recording in 2014 and information about the impact of COVID-19 on the data. DWP advises in the bulletins that annual changes are more reliable than quarterly registration figures to reflect longer-term trends.
c) The NINo Benefit Claimants bulletins are signposted through the landing page of these statistics, making it easier to access the statistics.
Based on the improvements DWP has made to the commentary in the NINo allocation report, we consider users are now better supported in the appropriate use of these statistics. The additional commentary makes it clearer for users to understand how these data fit into the wider migration data story, and the differences in sources.
The clear explanation of changes seen in the time series has added clarity and increased transparency.
The improvements in signposting to other related DWP NINo outputs offer greater accessibility to its data and statistics.
Requirement 2: To improve the value of NINo statistics DWP should establish a culture of wide, regular and informative engagement with a wide user community by:
a) Establishing effective engagement mechanisms to build open relationships with users outside government, as well as other relevant government departments, to allow direct consultation on their needs as users
b) Producing and publishing clear, explicit documentation explaining learning and responses from its user engagement, including who uses the statistics and what decisions they inform, and how it takes account of these users’ needs when planning publications
a) DWP has developed a User Engagement Strategy and an internal checklist for engagement. The strategy sets out the current engagement approach and future development plans for the publication alongside tentative timescales. The strategy includes a user engagement road map, an action plan of multiple strands – analytic, direct user engagement and research and development, and tentative timescales.
The team has strong links with policy teams within DWP and uses engagement opportunities with users and stakeholders outside DWP. It also has clear channels in which its users can contact them directly through the DWP website.
Team members from DWP present regularly at conferences (for example at the Royal Statistical Society conference in September 2020 alongside the Office for National Statistics (ONS) and the Migration Observatory).
b) Within DWP’s supporting information it states that identified users’ needs are reviewed and considered as part of ongoing development plans. Explicit documentation is not currently published on this although the key drivers for changes to the published statistics, as stated in its release strategy, include changing user needs and requests for data received through Parliamentary Questions (PQs) and Freedom of Information (FOI) requests.
DWP plans to publish more information about how users’ needs can and cannot be met in future.
DWP employs a number of approaches to engage with users. In particular the assessment team recognises that DWP is an active participant at migration statistics user events, collaborating with other statistics producers to remain visible and engage with key users.
We consider the User Engagement Strategy to be fairly comprehensive and it makes it clear that DWP wants to put users, and potential users, at the heart of their statistics.
These plans for engagement are very positive and we welcome DWP’s commitment to further engage with users outside government. This could involve DWP identifying potential users of the statistics, making better use of established user networks for migration statistics, and directly seeking feedback from users on how well they engage. Using this information to be able to understand the effectiveness of its engagement strategy should help to further strengthen user and stakeholder engagement over time.
The assessment team is happy to hear DWP’s plan to publish further information on how it responds to specific user needs or requests for data. This additional transparency will help to show how user view influence the teams planning and provide an additional route to explain to users how their needs are being considered.
Requirement 3: To improve the accessibility of the range of data available DWP should:
a) Outline its vision for the development of the NINo statistics, including its contribution to the cross-government analytical work using these data
b) Signpost the most recent and relevant ad hoc releases on the NINo data from the NINo webpages
c) Clearly set out what data are and are not available, including explanations about why certain data are not available, for example due to lack of data availability or lack of resources, to aid users’ understanding of what analyses can be carried out using NINo statistics
a) DWP has published a release strategy which details development plans for future releases with estimated timescales. For example, in 2021 DWP expects to extend its annual experimental statistics to include all DWP benefits and HMRC benefits.
DWP has a key role in the wider migration statistics transformation programme, as a member of the GSS steering group and on individual work streams. This is documented by ONS in its transformation updates.
b) DWP has added links to ad hoc releases on the NINo webpage, making these releases easier to find for users.
c) Information in Stat-Xplore outlines what data are available, to make it clear to users what analyses can be carried out.
We are pleased that DWP has taken positive steps in producing its own transparent communications on future developments through its release strategy.
The assessment team recognises DWP’s commitment and efforts to support ONS in contributing to the migration statistics transformation programme. DWP highlights to its NINo statistics users its wider role in the cross-government transformation programme and provides links to the ONS website for further details.
Quality
Requirement 4: DWP should assure itself and its users about the quality of the NINo statistics by:
a) Ensuring that it has a full understanding of the data quality assurance processes used in the production of the statistics, including the checks that are carried out by the data suppliers and the statistics team, and about any audits or third-party assurance of the source data
b) Making improvements to its processes necessary to assure users of good quality NINo statistics
c) Publishing a clear and comprehensive summary of its findings at an appropriate level of detail for its users
d) Publishing a process map to aid users’ understanding of the journey of the NINo data through the DWP and HMRC data systems and quality assurance processes undertaken at the various stages throughout the process
e) Ensuring that the administrative data sources used in the production of the NINo Benefit Claimants are listed in its SoAS. As part of meeting this Requirement, DWP should refer to the Authority’s Administrative Data Quality Assurance Toolkit.
a) DWP has published detailed Quality Assurance of Administrative Data (QAAD) supporting information in its methodology note. This outlines DWP’s data initial quality assurance checks of the Migrant Workers Scan data, the production and Quality Assurance (QA) of statistical data and products, peer reviewing of statistical products and sign-off by the Head of Profession for Statistics. The statistical team also has a QA checklist, which has been designed in line with OSR’s guidance, for auditing purposes.
b) As outlined in its quality document, the team has been working to automate the publication process and has been adopting the use of Reproducible Analytical Pipelines (RAP) to produce these statistics. Also, following a review of the full production and dissemination process in 2019, the statistics team streamlined the production code and QA processes to reduce the risk of human error while significantly reducing involved resource levels.
c) The team frequently reviews the commentary around key stories and trends and uses plain language and explains acronyms to cater for non-expert audiences. The statistical bulletin contains detailed and objective commentary on the current number and rates of change in NINo allocation to adult overseas nationals. Charts, visualisations and maps are used to illustrate key messages. The release strategy outlines that, in February 2020, additional context was added to better explain the trends in historic data.
d) DWP has published a NINo statistics production and quality assurance process map which documents the journey of the NINo data.
e) The Quality Report refers directly to using the Quality Assurance of Administrative Data guidance. The administrative data source used in the production of the statistics is clearly described in the methodology report.
DWP has undertaken considerable work to successfully meet this requirement. DWP has worked hard to improve its quality assurance documentation and its methodology note which informs users about the quality of the statistics. The approach to QA is clearly documented and the methodology note has brought significant clarity to these processes for users of the statistics.
The thorough and detailed information about the QA process contained in the publication reassures users of the quality of the statistics.
DWP’s use of RAP, and its dedication towards improving the quality of these statistics, is very positive.
The process map is a valuable addition to the publication. It clearly details the journey of the NINo data to provide insight into the data system, and reassurance of quality assurance processes, to users of the data.
The use of the QAAD guidance is an improvement. It is also made very clear what the data source is for the statistics.
Requirement 5: To enhance the public’s knowledge and understanding of the quality of NINo statistics and their appropriate use, DWP should publish a comprehensive and accessible set of quality and methods information that:
a) Outlines all the known limitations of the statistics, and the implications for the quality of the statistics, in terms of their existing use for measuring migration and other potential uses, and in relation to the European Statistical System (ESS) quality dimensions
b) Clearly explains the reasons for the discontinuities in the historical series, and how to account for quarterly seasonality when comparisons are made
c) Updates its published information to aid users’ understanding about the NINo application process, NINo eligibility, and the statistical methods used to produce the statistics, including the justification for the benefit hierarchy used in NINo Benefit Claimants
a) DWP has included extensive information about the limitations in the background information and methodology note, making it clear what the implications of the limitations are for the use of the statistics, in addition to providing links to the Code of Practice for Statistics and the European Code of Practice for Statistics.
b) The quality report and the bulletin provide information about discontinuities in the data. DWP explains that discontinuities are associated with operational changes. This also applies to the impact of COVID-19 on the statistics, which has been explained in detail. Further, DWP advises that annual comparisons should be made, rather than comparing on a quarterly basis, due to operational changes.
The background information provided has been developed to include information about NINo application, allocation and registration. Information is also included regarding eligibility for a NINo. A full and detailed explanation of the statistical methods used to the produce the statistics is also available, which provides detailed insight into this process.
We are pleased that DWP has expanded and improved the information about the limitations of the statistics in the background information and methodology webpage; this will help users to understand how the data can best be used and interpreted.
The information that has been included to aid users’ understanding of discontinuities is a positive addition to the publication. The information about the impact of COVID-19 is particularly detailed and will aid users in understanding the effect of this situation.
The inclusion of information about the NINo application process, eligibility, and the methods used to produce the statistics, has increased the quality of these statistics. DWP should seek to maintain the validity of this information for users, ensuring it is kept up to date.
Given the forthcoming changes to the migration system in UK, it will be extremely important for DWP to support appropriate use of its NINo statistics during this period. We expect DWP to offer users a similar level of support to aid users’ understanding of any limitations and discontinuities in the statistics.
Trustworthiness
Requirement 6: To enhance transparency in these statistics DWP should:
a) Publicly commit to announcing future changes to methods or classifications in advance of the publication of NINo Allocations
b) Provide sufficient explanation on the impact and magnitude of these changes on the use of the statistics
a) DWP has published a ‘Development Plans for Future Releases’ timetable for these statistics in the release strategy.
b) There have not been any recent significant changes made to the methodology but the ‘Development Plans for Future Releases’ timetable provides a mechanism by which users can be informed about the likely impact of any changes which will take place.
We consider the ‘Development Plans for Future Releases’ timetable to be a positive addition to the statistical publication.
DWP has committed to maintaining this development plan in future so users can be notified about planned changes. We look forward to DWP further developing its potential, for example, by including information about how it is responding to feedback from users.
Requirement 7: To increase the trustworthiness and to reassure users of the confidentiality of the NINo data DWP should publish documentation on the confidentiality arrangements that are in place. This might include information on encryption of data, memoranda of understanding with other government departments, and statistical disclosure control methods.DWP has published detailed information about the process of encrypting the data received from the Home Office in order to ensure confidentiality. Record-level data are encrypted and shared by the DWP data team and summary level data are shared under Memoranda of Understanding with other departments, all of which is outlined in the quality report. Further information about confidentiality is published in Stats-Xplore, which also explains the concept of introduced random error to further reduce the risk of identifiable data.The detailed information offers reassurance users about the confidentiality of the NINo data. There is a clear process DWP follow to ensure that data remains confidential at every step of the data journey.
Requirement 8:
To promote public confidence in the NINo statistics DWP should:
a) Reduce the number of people included on the NINo Allocations pre-release access list and provide a justification for each individual listed
b) Ensure that the pre-release access list is reviewed prior to each publication and updated accordingly, and includes the number of post holders where there is more than one
c) Include a link to the pre-release access list directly from its NINo webpage
a) DWP has reduced the number of people on the pre-release access list from 73 individuals to 17 named roles.
b) DWP has provided information about updates to the pre-release access list on their webpage which demonstrate that it is regularly reviewed. There is a clear process in place whereby the Head of Profession for Statistics reviews the list quarterly to ensure that access is still required by each person, and access is removed if this is no longer the case.
c) DWP has signposted its pre-release access list from the landing page of the statistics.
We are pleased to see that DWP has considerably reduced the number of people included on the pre-release access list and that there is a clear process in place to ensure this access is justified.
The signposting to the pre-release access list is a positive addition to the webpage.

 

Related Links:

Statistics on National Insurance Number Allocations to Adult Overseas Nationals

Stephen Balchin to Ed Humpherson: Assessment of DWP statistics on national insurance numbers (NINo) allocated to adult overseas nationals and access to work

Ed Humpherson to Stephen Balchin: Statistics on national insurance number allocations to adult overseas nationals 

 

Ed Humpherson to Peter Schofield: Universal Credit management information

Dear Peter

UNIVERSAL CREDIT MANAGEMENT INFORMATION

Thank you for your response to our letter dated 22 April, which sets out the future publication of information on Universal Credit. Our understanding is that you will be publishing a weekly management information series summarising declarations made by individuals and households to Universal Credit and information on advance payments of Universal Credit. Data on geographical breakdowns and other characteristics will continue to be published as part of the monthly Universal Credit Official Statistics.

It is reassuring that you recognise the importance of not using unpublished information in public statements and we hope that the weekly series mitigates the risk of this issue reoccurring. We are aware that the timing of the work and pensions oral session is not aligned with the release of weekly management information and therefore Ministers may occasionally wish to refer to additional data. In these cases, you should look to ensure this additional data is made publicly available to support the statement being made.

We will continue to monitor the evolving interest in data on Universal Credit and work with your department to ensure its statistics remain aligned to the Code of Practice for Statistics

I am copying this letter to Steve Ellerd-Elliott, Chief Statistician for the Department for Work and Pensions.

Yours sincerely

Ed Humpherson

Director General for Regulation

 

Related links

Ed Humpherson to Steve Ellerd-Elliot: Universal Credit management information

Response from Peter Schofield to Ed Humpherson: Universal Credit management information

Response from Peter Schofield to Ed Humpherson: Universal Credit management information

Dear Ed,

UNIVERSAL CREDIT MANAGEMENT INFORMATION

Thank you for your letter of 22 April. I was pleased to read your positive comments about the supplementary information we have published alongside our Official Statistics but I equally recognise your concerns around the use of unpublished information and that the release did not include all of the information that we had preannounced.

I take the matter of handling and using statistics very seriously and I am sorry that on this occasion there were issues in the way in which we have handled this release. This is an exceptional release which sees us publish more timely information than we ever have before on Universal Credit to meet the needs of users and our duty of equality of access at this challenging time.

The release was developed at speed and, whilst we were working with the best intentions, it was a mistake that the pre-announcement was made whilst internal discussions were ongoing and before agreement had been reached on what additional data we would be able to release. I recognise that on this occasion it was not accurate on the information that would be published.

Our Chief Statistician, Steve Ellerd-Elliott, has reviewed what happened and is updating our guidance and processes to ensure appropriate approval and sign-off is in place for this type of release in future.

We will continue to provide the full Universal Credit breakdown in our monthly Official Statistics but during this unprecedented time when Universal Credit claims are at an exceptionally high level we also want to provide more timely information. I am mindful of the need to do this in a way which ensures that any public statements are supported by information which is equally available to all.

The management information we have published around claims for Universal Credit and advances are matters that have been raised by the Work and Pensions Committee and that have been used in the public updates we have given on the Department’s performance. We will now be increasing the frequency to publish this on a weekly basis. The first release will be published on 5 May and then weekly until at least the end of June.

We will continue to review the information available on Universal Credit and other benefits, and aim to provide informative statistics that are trustworthy, high quality and represent good value.

I welcome your organisation’s continuing positive work with the Department and I am happy to discuss ways in which the Department can continue to improve its release of statistics.

I have copied this letter to Steve Ellerd-Elliott, DWP Chief Statistician and will also send a copy to the WPSC.

Yours sincerely,
Peter Schofield CB

 

Related links:

Ed Humpherson to Steve Ellerd-Elliot: Universal Credit management information

Ed Humpherson to Peter Schofield: Universal Credit management information

Ed Humpherson to Steve Ellerd-Elliott: Universal Credit management information

Dear Steve

UNIVERSAL CREDIT MANAGEMENT INFORMATION

 I am writing to welcome the publication of supplementary management information on Universal Credit yesterday by the Department for Work and Pensions (DWP), alongside the regular Official Statistics release, but also to share our concerns about how this differs from the information that was preannounced. Universal Credit continues to be an area of high user interest and has increasingly moved to the foreground of public debate as the UK adjusts to rapid changes in society and the economy as a result of Covid-19.

We recognise the challenging environment in which your department is working in and support the work it is doing to balance the demand for up-to-date information with ensuring relevance to users and good quality data. The statistics teams in your department have shown flexibility and responsiveness by exploring different data sources to meet this demand.

In recent weeks, this management information has been provided by ministers to give context in public statements. Following yesterday’s publication, and in line with the Code of Practice for Statistics, the DWP should ensure that any public statements are supported by information which is equally available to all.

It is both regrettable and concerning that some of the breakdowns of management information that your department had preannounced last week and raised with the Work and Pensions Committee were not published yesterday. On 15 April 2020, the DWP preannounced the release of supplementary management information and stated it would include breakdowns by geography and key characteristics. Prior to this Peter Schofield, Permanent Secretary of the DWP, wrote to the chair of the Work and Pensions Committee on the 3rd April 2020 stating the DWP was expecting to publish a set of supplementary data on Universal Credit including the number of declarations, the number of awards, the number of advances and information relating to the busiest times of the day for calls and other relevant performance information. Yesterday, however, the DWP only published information on the number of households and individuals making a Universal Credit declaration and the number of Universal Credit advances by type of advance.

The labour market will continue to feel the effects of Covid-19 over the coming months and therefore user demand for the timely, relevant data this management information provides will probably persist. We hope that the DWP will publish the remaining breakdowns that it had preannounced and we look forward to seeing this information published on a regular basis for the duration it remains relevant to users. It is our understanding that the DWP had already been working on these breakdowns. As management information produced by the DWP is subject to voluntary application of the Code of Practice for Statistics, the DWP should look to enhance its trustworthiness by publishing information it had committed publicly to release. Not doing so creates a risk to both transparency and to public understanding.

The DWP should be transparent about the frequency it intends to publish this management information once it has been determined. We will keep in touch with the statistics team as part of our ongoing monitoring of statistics on Universal Credit.

I am copying this letter to Peter Schofield, Permanent Secretary of the DWP.

Yours sincerely

Ed Humpherson

Director General for Regulation

Compliance Check of DWP Universal Credit Statistics

Dear Steve

UNIVERSAL CREDIT STATISTICS

As you are aware, we recently completed a compliance check of the DWP Universal Credit (UC) statistics. We considered the trustworthiness, quality and value of these statistics in relation to the Code of Practice for Statistics. We have made a number of recommendations to support your continued development of these important experimental statistics.

We appreciate the positive way that the team has engaged with us during the compliance check, and it is clearly committed to delivering improvements. Our Labour Market and Welfare Domain Lead, Catherine Bremner, will continue to engage with the team on progress over the coming months.

We found a range of positive features that demonstrate the value and quality of the statistics. The statistics are easily accessible, and the gov.uk landing page is well laid out, with useful links to previous publications, publication release strategy and interactive maps and data tools. It also highlights related publications, research and analysis papers, evaluation frameworks, and ad hoc statistical publications, all of which add insight and value. The bulletin is easy to follow, the ‘What you need to know’ and ‘About these statistics’ sections explain the key features of the statistics, and informative subheadings explain key trends. We welcome the comprehensive report on the limitations of the data sources and the clear definitions in the Background and Methodology document.

We identified several areas for improvement which would enhance the clarity, value and quality of these statistics. We consider that the commentary in the bulletin could generate more insight for instance by explaining what the different comparisons mean, such as those between male and female claimants. We also encourage you to consider more-informative ways of presenting the data, for example a trend analysis of the four measures (claims made, starts, people on Universal Credit, households on Universal Credit).

We encourage the team to explain technical terms to help all users understand key concepts and methods, such as ‘pathfinder’ and ‘conditionality regime’. In addition, it may be helpful for you to review whether some of the visualisations could be improved to aid interpretation of the statistics; for instance, the pie charts or the map with the breakdown on UC claimants by local authority, which lacks context. The team may find the Good Practice Team’s guidance on effective charts and maps helpful.

We welcome the section on uses and users of the statistics and would look to DWP to proactively engage with external users and stakeholders to encourage wider use of the statistics, and to ensure that users’ needs are understood, and can feed into further developments of the statistics.

To help users understand the limitations of the statistics and the methods, further quality information could be included on the impact of the limitations, and information on methods changes and why the four measures (claims, starts, people, households) were chosen. Further information on comparability with other statistics would help users generate additional insight, by explaining why these comparisons are relevant and important. It would also be useful to signpost users to related statistics and explain why they are relevant, for example, the Northern Ireland Universal Credit statistics produced by the Department for Communities.

An essential part of assuring yourselves and users about quality, and enhancing the trustworthiness of the statistics, is to provide information about quality assurance. Producers of statistics should explain clearly how the statistics and data are accurate, reliable, coherent and timely. As these statistics are based on administrative data, we encourage the team to apply our Quality Assurance of Administrative Data (QAAD) framework to assure users about the quality assurance arrangements and to help them understand how the Universal Credit data are collected and processed.

These statistics have been published as experimental statistics since they were first introduced in December 2013. We encourage the team to continue to develop these important statistics, to enhance public confidence in their trustworthiness, quality and value. In general, we feel that you could be more ambitious about improvements to the Universal Credit statistics, focusing on how they can enhance value through, for example, linkage with other DWP benefit data and survey data.

Please do not hesitate to contact us if you would like to discuss this further.

Yours sincerely

Mark Pont
Assessment Programme Lead

Compliance Check of DWP Benefit Cap Statistics

Dear Steve

BENEFIT CAP STATISTICS

As you are aware, we recently completed a compliance check of the DWP Benefit Cap statistics. We considered the trustworthiness, quality and value of these statistics in relation to the Code of Practice for Statistics. We have made a number of recommendations to support your continued development of these important experimental statistics.

We appreciate the positive way that the team has engaged with us during the compliance check and it was encouraging to hear of its development plans and desire to maintain contact and discuss improvements. Our Labour Market and Welfare Domain Lead, Catherine Bremner, will continue to engage with the team on progress over the coming months.

We found many examples of good practice. The background information on the statistics is clear and informative, providing a good description of what Benefit Cap is and how it works. The background and methodology document also explains the policy context and contains links to pages on the key benefits. The statistics are easily accessible and the gov.uk landing page is easy to follow, with useful links to previous publications, interactive tools, related publications and research papers, all of which help users better understand the statistics and data. We identified the need to present a clear and coherent narrative about benefit statistics in the bulletin, that is easy to read and supports use by all types of users.

We welcome the helpful summary of key points about the methods at the start of the bulletin and the detailed explanation on the limitations of the statistics in the background and methodology document. The reasons for change and their impact on the statistics are helpfully explained in depth along with how the outcome categories have changed. We have identified a range of detailed improvements that we consider could be made to the presentation of the statistics, which we have passed on to the team separately.

We are encouraged by the recent development of a regular user group and engagement with local authorities, along with embedding user engagement as part of the production process.

We would encourage more engagement with external users and stakeholders to promote wider use of the statistics, to ensure user needs are understood and to maximise public value.

An essential part of assuring yourselves and users about quality, and enhancing the trustworthiness of the statistics, is to provide information about quality assurance. Producers of statistics should explain clearly how the statistics and data are accurate, reliable, coherent and timely. As these statistics are based on administrative data, the team has agreed to produce a quality report using our Quality Assurance of Administrative Data (QAAD) framework to assure users about the quality assurance arrangements and to help them understand how the Benefit Cap data are collected and processed. This is a good development.

Please do not hesitate to contact us if you would like to discuss this further.

Yours sincerely

Mark Pont
Assessment Programme Lead

Statements on ‘new jobs’

Dear Steve

Thank you for responding to my query regarding an article published in the Daily Telegraph on the 21st April 2019, in which the Minister of State for Employment, Alok Sharma MP, was discussing UK employment.

The original article made several statements, such as that “Britons have filled nearly all of the new jobs created in the UK since the 2016 referendum” and that “UK workers have accounted for around nine in ten of new people in work since 2016, compared to half of the people entering work in the two years before”.

You recognised that these statements are not consistent with the relevant official statistics. You explained that the Department for Work and Pension’s internal analytical briefing discourages the use of language which confuses changes in employment with the different and vaguer concept of “new jobs”, in particularly avoiding:

  • The use of the word “job(s)” when referring to headline Office for National Statistics measures of the number of people in employment.
  • The use of the phrase ‘job creation’ where it is claimed or implied that the Government is responsible for directly creating or is the main reason for the creation of “new jobs”.
  • The mixing of statistics on jobs and employment to avoid public confusion.
  • Anything that may confuse net changes in employment with total flows into employment.

This briefing is important and helpful. On this basis, I am reassured that this is an isolated case in which the Department’s analytical advice has not been followed in full. I am also pleased to note that the Telegraph article has been amended and all references to “job creation” and “new people in work” have been removed.

Yours sincerely

Ed Humpherson
Director General for Regulation