Members present

• Jonathan Portes (Chair)
• David Bell (University of Stirling)
• Alexandra Fitzpatrick (Welsh Government)
• Chaitra Nagaraja (University of Exeter)
• Nye Cominetti (The Resolution Foundation)
• Tom Pybus (HM Treasury, Delegate for Daniel Gallagher)
• Richard Murray (Scottish Government)
• Tim Butcher (Low Pay Commission)
• Jonathan Wadsworth (Royal Holloway, University of London)

Labour Market Technical Group members

• Alice Crundwell (Bank of England)
• Douglas Rendle (Bank of England)
• Michal Stelmach (Bank of England)
• Paul Cheston (HM Treasury)
• Sunita Bali (Office for Budget Responsibility)
• Edward Robson (Office for Budget Responsibility)
• John Ferrier (Department for Work and Pension)
• Andrew Ward (Department for Work and Pensions)
• Ivan Bishop (Department for Business and Trade)

Northern Ireland Statistics and Research Agency (NISRA) presenters

• Philip Wales
• Brian Green

NISRA colleagues

• Gerard Colgan
• Sandy Fitzpatrick
• Gillian McCrory
• Brendan Morgan

Secretariat

• Sarah Ash

ONS colleagues

• Liz McKeown
• Heather Bovill
• Gareth Evans
• Sian Evans
• Tom Evans
• James Harris
• Orlaith Fraser
• Colin Beavan-Seymour
• Debbie Cooper
• Sarah Henry
• Myrto Miltiadou
• Katy Nicholls
• Skevi Pericleous
• Miranda Winchcombe-Young

1. Introduction

1. Jonathan Portes opened the meeting and invited Liz McKeown to provide an update on the publication of the Devereux Review of the ONS and the Plan for ONS Economic Statistics.
2. Liz McKeown presented on the newly published ONS Plan for Economic Statistics, highlighting improvements to core statistics, survey systems, and quality assurance. She emphasised the importance of continuous improvement and engagement with users.

2. Presentation on NISRA’s Transformation of the NI Labour Force Survey – Philip Wales and Brian Green

1. Philip Wales (PW) opened the session by thanking the panel for the opportunity to present NISRA’s work. He noted that the presentation would cover the design philosophy, survey implementation, and early results from the new Northern Ireland Labour Market Survey (LMS), which in time will replace the traditional Labour Force Survey (LFS) in NI.
2. PW highlighted:
   • The aim was to gather feedback from panel members on:
     • The design and trade-offs made in the survey.
     • The results and how they compare to benchmarks.
     • The steps being taken to understand differences between LMS and LFS.
   • He emphasized the importance of user engagement and the opportunity to shape the future of labour market statistics in NI.
3. Brian Green (BG) presented on the LMS. BG first spoke of NI LMS Survey Design and Implementation:
   • Launch & Pilot: LMS was piloted in May 2023 and launched in October 2023.
   • Sample Size: 170 households per week (2,210 per quarter), a 70% increase over the NI LFS.
   • Methodology: Online-first with follow-up by field staff from day 11. Fieldwork lasts 28 days per wave; all non-responding households at Wave 1 are followed up by field staff and all Wave 1 households are invited to take part in Waves 2–5.
   • Communications: Tailored, branded materials (postcards, letters) with QR codes and Unique Access Codes. £10 incentive offered to encourage participation at each Wave.
4. NI LMS Questionnaire Design:
   • Duration: Targeted under 20 minutes; Wave 1 averages 15 minutes, Wave 2 around 12 minutes.
   • Modular Design: Avoids repeating unchanged information across waves.
   • Wave 2 Module: Focus on work quality (e.g. career progression, meaningful work, managerial support, involvement in decision making, whether skills match the job, whether they have been bullied or harassed, flexible working arrangements and satisfaction with their job). There is scope to add further modules across Waves 3-5.
   • Content Adjustments:
     • Removed questions with poor data quality or available from other sources (e.g. income, smoking and health, well-being, accidents at work).
     • Plans to link LMS data with administrative sources for enhanced outputs and quality assurance.
5. NI LMS Survey Performance:
   • Response Rates:
     • Wave 1: 59.4% (vs. 50% for LFS).
     • 96% of responses are fully complete.
     • 64% online responses; 36% involve interviewer support.
     • Retention across waves exceeds 80%.
   • Representativeness:
     • Slight underrepresentation of younger age groups; overrepresentation of 65+.
     • Weighting methodology effectively adjusts for demographic and geographic imbalances.
     • Deprivation-based response rates range from 47.5% (most deprived) to 65.6% (least deprived).
6. NI Supporting Projects:
   • Crossover Survey: 410 LFS households invited to complete LMS; 76% response rate so far.
   • Data Linkage:
     • LMS and LFS to be linked with Universal Credit and PAYE data.
     • Aim to improve admin based unemployment and employment comparator indicators.
   • Derived Variables (DVs):
     • 69–70 DVs used; peer review requested for 5 key ones.
   • Processing System:
     • Built in-house by NISRA with ONS support, based on census processing.
     • Data aimed to be delivered to the ONS within 6 working days post-fieldwork, for ONS to apply seasonal adjustment at UK level.
7. NI Next Steps:
   • Finalise NI data processing systems
   • Peer review of key NI derived variables
   • Analyse results from NI Crossover survey (completes 29^th June 25)
   • Expand analysis and documentation.
     • Disability
     • NI Inactivity rate
   • Continue collaboration with ONS on UK-level outputs solution and discontinuity adjustments
   • Plan transition from LFS to LMS (timing TBD).

3. Panel Discussion

1. Jonathan Portes (Chair)
   1. Praised the clarity and comprehensiveness of the presentation.
   2. Emphasized the importance of short, easy-to-complete questionnaires.
   3. Raised concerns about comparability with the rest of the UK, especially regarding non-response bias in unweighted categories like disability.
2. The panel discussed:
   1. Differential response rates by religion, given its relevance in NI.
   2. The importance of linking LMS to earnings data (e.g. ASHE) for religious group analysis.
   3. The large differences in inactivity rates between LMS and LFS, requiring further investigation.
3. BG and PW responded:
   1. Acknowledged the need for deeper analysis, especially on disability and religion.
   2. Confirmed early signs of narrowing response gaps between nationalist and unionist areas.
   3. Suggested that higher employment among younger people may explain lower inactivity in LMS.
   4. Noted that online surveys may better capture certain groups (e.g. young part-time workers).
   5. Reiterated plans to link LMS with administrative data to validate findings.
4. BG and PW spoke on Earnings Data:
   1. Decision to exclude earnings from LMS was based on concerns about data quality and respondent burden.
   2. May revisit inclusion in Waves 3–5, acknowledged the importance especially for self-employed, where RTI data is lacking.
   3. Aim to derive hourly earnings by linking LMS hours data with RTI pay data.
5. BG and PW also commented on sharing results:
   1. Agreed to share slides (excluding sensitive results).
   2. Committed to providing follow-up analysis on inactivity and earnings for DWP and Low Pay Commission.

4. Conclusion/Next Steps

1. Next steps are:
   1. Continued analysis of LMS data and comparison with LFS.
   2. Peer review of derived variables and processing systems.
   3. Ongoing collaboration with ONS on UK-level outputs and methodology.
   4. Further user engagement and documentation development.
2. The next regular meeting of the Stakeholder Advisory Panel is scheduled for 16 July 2025.