The data gathered by large economic surveys remain a vital component of many key Office for National Statistics figures, especially those on the labour market. Maintaining good survey response rates is a challenge faced by national statistical institutes worldwide. In this post Darren Morgan sets out how the ONS is addressing this issue and tapping alternative sources for the important information we need.
Measuring the labour market
In the UK we have three key metrics for monitoring the world of work: our Workforce Jobs survey primarily asks employers across different sectors how many people they are employing. We also now have a powerful source of administrative data in the shape of HMRC’s Real Time Information from PAYE (RTI). This shows how many people are in work and paying tax in a given period.
These two sources have shown similar patterns of how many people have been employees during and post the pandemic. However, they tell us nothing about those not working, or why they are not working. At a time when vacancies remain well above pre-pandemic levels, it’s crucial that we find out why people are out of work.
That’s where the regular – and still very large – ONS Labour Force Survey comes in. It asks a sample of tens of thousands of people whether they are working and if not, why not.
For example, thanks to the Labour Force Survey, we know that a significant number of older workers retired early at the start of the pandemic, many of whom are now coming back, while an increasing number of people of all ages are suffering from long-term sickness.
New challenges
However, in common with national statistical institutes worldwide, we are finding it a challenge to maintain response rates – particularly among those who are in work and, it’s fair to assume, have less time to take part in surveys. Meanwhile the Labour Force Survey’s headline numbers and trends for employment are starting to vary from Workforce Jobs and our RTI data, while lower response rates could also be leading to increased sampling variability for unemployment and inactivity too.
Responding to the challenge
To improve response rates, we are bringing in improvements to the Labour Force Survey, asking more people in different ways about their engagement with the jobs market. The first figures from this new transformed survey are due to be published in our headline release next spring.
As well as introducing new collection methods and asking more people, we are reweighting the existing Labour Force Survey to take account of the number and type of people responding to the survey. In October we will introduce much more up-to-date population information to make sure the survey responses are matched to current population demographics. But we are also assessing what other metrics can be used – such as, potentially, Self Assessment, Claimant Count or RTI – to adjust the raw survey responses in future to ensure published data are as accurate as possible.
But in the meantime, to get a clearer picture of changes in the labour market it is crucial to draw on a number of sources such as Workforce Jobs, RTI and the Labour Force Survey.
Encouraging people to give us the information we need to produce key economic data is clearly a challenge, but here at the ONS we are working hard to ensure our labour market data continue to accurately reflect changes in our complex and evolving economy.
Editor’s note: since this blog was published on 15 August 2023, the ONS has decided not to proceed with reweighting the LFS in October (see the notice on our releases); instead the latest population estimates will be taken on board alongside the introduction of the transformed survey in spring 2024.
Darren Morgan is Director of Economic Statistics Production and Analysis