Recently there has been coverage of the challenges facing certain Office for National Statistics (ONS) household surveys. Due to concerns about the quality of data resulting from a particularly low response rate, In October 2023, the ONS suspended the publication of one of its headline labour market surveys. This blog sets out some of the implications for labour market data in Scotland.
Employment statistics in the UK
Declining response rates to key surveys is not a new challenge, nor one unique to the UK. The Financial Times noted that the response rate to the ONS Labour Force Survey has declined from around 50 per cent to around 15 per cent over the last ten years, which has led to the data no longer meeting the necessary quality standards to be designated as ‘national statistics’ (which denotes that the data meets the highest statistical standards in the UK). In August 2023 the collection of data for the Labour Force Survey faced a particularly acute challenge. The Labour Force Survey is one of three sources of data the ONS use to monitor the labour market.
- The Labour Force Survey is the largest household study in the UK, and provides the official measures of employment and unemployment. It surveys around 40,000 households UK-wide every quarter, and is the only regular source of information on self-employment (around 290,000 people in Scotland).
- Workforce jobs is a quarterly publication which draws from 2 surveys of employers (covering the public and private sector), and the Labour Force Survey to measure short term employment by industry, region, gender, and full or part time status.
- Access to HMRC Real Time Information from the Pay as you earn (PAYE) taxation system. This system gathers information for HMRC on PAYE deductions every time an employee is paid, and has been required for all employers since April 2014. HMRC note that this currently covers 48 million employees.
Confidence intervals and the tendency to overstate accuracy
There is a trade off between accuracy and timeliness in producing statistics. When official statistics are compiled based on survey data, rather than administrative data, the headline figures are not precise, but are rather subject to confidence intervals. The Fraser of Allander Institute have explained this in several publications:
“For example, the 95% ‘confidence interval’ for Scotland’s unemployment rate – i.e. the range within which we can be 95% certain that the ‘true’ value lies within – is ± 0.8 percentage points. So in other words, the current unemployment rate in Scotland could lie anywhere between 5.7% and 4.1%!”
The declining response rate make this uncertainty around the headline rate greater. To illustrate this point, the charts below set out the headline measures of unemployment in Scotland from the last four labour market trends publications based on the Labour Force Survey, which cover quarterly figures for February to April 2023, to May to July 2023. The chart sets out the headline figure as presented as a single percentage, and also shows the confidence intervals around each figure. It’s worth noting that the declining response rate can be seen in the growth of the confidence intervals. Initially it’s plus or minus 0.9 percentage points, but in the most recent data it has grown to plus or minus 1.3 percentage points.
How are the ONS responding to these challenges?
As noted above, the response rate to the Labour Force Survey has been declining for some time, and the ONS have been working to explore how the collection can be improved. The ONS have been running a ‘transformed Labour Force Survey’ (TLFS) concurrently with the LFS throughout 2023, which is intended in time to replace the LFS. The LFS, prior to the pandemic, was conducted by writing to UK Households and following up with visits for those who did not respond, but during the pandemic when this was no longer possible the in-person visits were replaced with phone calls. The ONS note that this switch:
“was leading to bias in responses, with particular demographics such as the older population more likely to respond, while we were less likely to hear from people who rented their property.”
The TLFS will increase the sample size and change the way data is collected. This will be an ‘online first’ approach, supported by telephone collections and a ‘knock to nudge’ follow up to encourage participation.
Further updates were provided in an article on 2 November, setting out how the ONS plan to reintroduce LFS estimates and progress on the implementation of TLFS. This plan involves:
- Reintroducing in-home interviewing as part of the LFS
- Increased contacting of households who do not respond, which is expected to lead to an additional 1,200 visits per week
- Increasing the sample size from 16,000 to 24,000
- Recruiting additional field staff to improve data collection
The ONS aim to replace the LFS with the TLFS from March 2024, subject to final data quality checks. The most recent publication of labour market data on 14 November included:
- Labour market overview, UK bulletin
- Average weekly earnings in Great Britain bulletin
- Vacancies and jobs in the UK bulletin
- X10: Adjusted employment, unemployment, and economic inactivity dataset (seasonally adjusted, including regional and age breakdowns)
However, the November publication did not include:
- Employment in the UK bulletin
- Labour market in the regions of the UK bulletin
- Quarterly datasets, which include labour market flows
- Young people not in education, employment or training (NEET) bulletin
- Working and workless households in the UK bulletin
In October, the ONS published the X10 experimental data which takes data from the April to June LFS returns and adjusts it using administrative data from HMRC and claimant count data. The October publication included only a breakdown by age bracket, and November’s added a regional breakdown to this.
On 14 November, the ONS published a review of its experimental labour market statistics. This review sets out the detailed context for the challenges facing the Labour Force Survey, and assesses the suitability of the revised methodology and quality assurance to produce the headline labour market estimates.
Alternative sources of labour market data
The Bank of England is a high-profile user of ONS data, which helps to inform decisions by the Monetary Policy Committee on interest rates and other policy decisions. The labour market can provide an important measure of economic activity, and therefore possible paths for domestic inflation. The Economist highlight the importance of a good understanding of the labour market for policy makers, noting that:
“Policymakers fret that despite a slowing economy, the jobs market remains tight and rapid wage growth could entrench high inflation. One problem the bank faces is that core jobs-market data are now much less reliable than they once were. The speedometer may be faulty.”
In the Bank of England’s November Monetary Policy Report, the Bank notes that it considers a number of data sources to inform its view of the labour market. Discussing the different indicators for the labour market in more detail, the Bank states that:
“The most recent ONS Workforce Jobs and HMRC payrolls data point to a modest but positive quarterly growth in employment in Q3, and a small contraction in employment growth in Q4. There have been notable uncertainties surrounding recent Labour Force Survey (LFS) estimates. These have made the data harder to interpret and have resulted in the ONS temporarily pausing its publication of LFS estimates following the June data (Box B). Alternative experimental statistics published by the ONS, which take the LFS employment estimate in the three months to June and project it forward in line with the HMRC payrolls data thereafter, suggest that employment fell by 0.2% in the three months to August. As experimental statistics, these estimates need to be interpreted with caution.”
Scottish Government labour market publications
Until August 2022, the Scottish Government produced two monthly labour market publications sourced from ONS data.
- Labour market trends covers the latest estimates of employment, unemployment and economic inactivity from the Labour Force Survey, as well as experimental data from HMRC’s pay as you earn real time information system.
- The Labour market monthly briefing covered headline estimates, vacancies, the disability and ethnicity employment gaps, unemployment and the claimant count. Since August 2022, this has no longer been published.
In October 2023, the Scottish Government cancelled the October publication of Labour market trends, and explained that:
“This publication has been cancelled following the ONS operational note outlining there will be no release of ONS Labour Force Survey estimates for June to August 2023 for Scotland and the other countries and regions of the UK”
In November, the Scottish Government published Labour Market trends using the experimental data produced by the ONS. The Scottish Government note that these experimental statistics are classified as ‘Official Statistics in Development’, which will be tested with users against the values set out in the Code of Practice for Statistics.
Challenges around the reliability of data on the Scottish labour market pose several questions for policy makers
The Scottish Fiscal Commission, in their Fiscal Sustainability Report, outlined the demographic challenges that Scotland faces and the impact this will have on public finances. Across the developed world, working age populations are growing more slowly than the population as a whole, reducing the number of taxpayers to fund public services while increasing demand (this is referred to as an increasing dependency ratio). Scotland’s dependency ratio is expected to increase more quickly than the UK as a whole. The Scottish Government’s 10 year economic strategy pledges to “Systemically address Scotland’s labour market inactivity challenges” which will be delivered through:
“Assess[ing] trends within different labour market inactive groups and understand what steps can be taken to bring more individuals into the labour market – including through the use of childcare and transport provision, part-time/flexible working, support for employees with disabilities, and business start-up and work from home opportunities.”
For policy to be effective, good quality and timely data will be required in order to understand the scale of inactivity and crucially the barriers people face in accessing the labour market.
In the shorter term, the Scottish Government have pledged to spend £500 million over ten years through their Just Transition Fund to ensure that workers in sectors which rely on high carbon industries are not left behind in the transition to a net zero economy. Ensuring that funds are deployed in the right areas and at the right time will require a good understanding of sectoral and regional employment trends.
Andrew Feeney-Seale, Senior Researcher, Financial Scrutiny Unit