Dear Lord Shipley,

I write in response to the Youth Unemployment Committee’s report, “skills for every young person”. I am also following up on the Office for National Statistics (ONS) ongoing work to measure youth labour market status against economic background.

Recommendation: The Government must work with the ONS to improve the quality and quantity of employment data collected on specific groups of young people, in particular those from disadvantaged (such as FSM-eligible) and ethnic minority backgrounds. This data must be published at more regular intervals than is presently the case so that it can be interrogated by policymakers.

The ONS currently publishes data on employment status by ethnicity and by disability every quarter. These data are available from the Labour Force Survey (LFS), a quarterly run household survey of the employment circumstances of the UK population. Although specific statistics of young people could be produced from this source, due to sample sizes, they would not be of sufficient quality to reliably measure the labour market conditions for this level of granularity.

Currently we use an expanded version of the LFS, the Annual Population Survey (APS) for such granular estimates. This survey uses a larger sample and is available for 12-month periods four times a year i.e. January to December, April to March, July to June and October to September. Again the limiting factor would be sample size which affects the quality of very granular analysis.

To improve the granularity of these data, we are currently developing an “online first” survey called the Labour Market Survey (LMS). The LMS is a mixed-mode survey that focuses on the core data collection requirements needed to produce labour market estimates. As the LMS is designed to have a larger sample size it will improve the quality and quantity of estimates available for specific groups of young people. It should be noted however, that given the potential number of characteristics and the small size of some of the population groups, we will still need to be mindful of statistical quality where analysis becomes particularly granular.

Additionally, the increased sample size of the LMS will give us options around the balance between regularity and quality of our estimates. We will work with key users to take advantage of those options to tailor our output strategy to best suit their needs.

Measuring youth labour market status against economic background

As outlined in my letter from October 2021, I promised to follow up with the Committee on the measurement of youth unemployment by economic background. The Committee were interested in knowing whether we could produce analysis on this using the ONS LFS.

We have assessed whether this is possible and have concluded that it is.

We have looked at the social mobility data in the LFS, which are collected annually between July and September. These data collect information on an individual’s socio-economic background when they were 14 years old:

  • Where the respondent lived
  • Household composition (i.e., with parents, with other family, not living with family)
  • Main wage earner in the household
  • Occupation of main wage earner (i.e., a parent/guardian, joint-earners, or no earners)
  • Whether main wage earner is an employee or self-employed

We have assessed the possibility of producing analysis using these data, and their quality. Our main concern was small counts, given the small sample sizes of the young unemployed when broken down by socioeconomic background. We have looked at individual counts, non-responses, and proxy responses. We have investigated potential issues specific to the young, such as, likelihood of being unemployed or underemployed at some point in their life. We have also assessed whether we can breakdown the data by protected characteristics or by other characteristics of interest (e.g., regions).

We have concluded that the best way to present a breakdown of labour market status by economic background is to produce a measure of socioeconomic background using the National Statistics Socio-Economic Classification, (NS-SEC),  using occupation. Using this method, we can derive the NS-SEC of an individual when they were 14 years old using the occupation of the main wage earner in the household at that time. This also covers individuals living in households where no one is in work. We believe that this is the best measure of socioeconomic background we can produce using the social mobility data in the LFS. The NS-SEC is considered to be one of the most accurate measures of socioeconomic status and is widely used in academia.

We will analyse both the unemployed and those who are not in work (unemployed and economically inactive) when it is possible and there are no sample size limitations.

We will present these data on the young people ages 16 to 24 broken down by smaller age groups when possible, i.e. 16-17, 18-21 and 22-24. This is because these three age groups are at a very different stage of their working life: the 16–17 age group are in mandatory education, the 18–21 age group are either in higher education or entering the labour market, while most young people in the 22–24 age group are already in the labour market. We will look at sex and disability, but not other protected characteristics due to the small sample. We also plan to compare our results to other age groups, for example those aged 25-34 or 35-49.

We expect to produce an article summarising our proposed analysis and findings by the end of April 2022, which will be shared with the Committee.

Please do not hesitate to contact me if I can be of any further assistance.

Yours sincerely,

Darren Morgan

Director, Economic Statistics Production and Analysis