Measuring the labour market during Coronavirus
The coronavirus is clearly having a very serious effect on the UK economy. Measuring its impact on the world of work presents unique challenges. Ahead of the latest labour market statistics release, David Freeman looks at how the ONS is responding to the pressing need for new information.
COVID-19 is having not only a big impact on the labour market but also how we measure it. The effects on households and businesses from sickness, self-isolation, homeworking and the provision of government support all need to be captured, and as quickly as possible. This means adapting and improving many of our sources and methods.
Providing the numbers people need
While the headline figures on jobs, unemployment and pay are still important, many people also want information on the specific impacts of COVID-19, such as changes in the level of homeworking and the impact of the employee furloughing scheme. To meet these demands, we have quickly launched a new online survey, which is already providing data and analysis on the impact of COVID-19. The Business Impact of Coronavirus Survey (BICS) is a fortnightly survey of businesses and includes effects on the workforce, such as working at home, furloughing of staff and changes in the number of employees.
In addition, the ONS has already published various articles on the impact of COVID-19. Several of these cover topics relevant to the labour market, including self-employment, the financial resilience of households when employment income drops, working parents, older workers and homeworking. We have also added questions to our existing surveys as well as quickly developing another new one with an online-only questionnaire:
- New questions on the Labour Force Survey (LFS) – from April, the LFS has asked additional questions relating to the impact of COVID-19 and the support schemes relating to employees and the self-employed.
- Labour Market Survey (LMS) – an on-line survey of households that went live at the end of March. This collects key labour market information as well as some specific questions around the impact of COVID-19.
Closing the reporting gap
Our regular publication cycle means that we publish the majority of figures on work and earnings around six to seven weeks after the end of the reference period – so, for example this month (April) we will be publishing LFS results for the December-February period. This normally gives a reasonable balance between timeliness on the one hand and ensuring that we have sufficient survey responses and enough time to check the data on the other.
However, in the current situation, it is even more important for policymakers to know what is happening now. In recent months the ONS and HM Revenue and Customs have jointly published estimates of employment and earnings based on tax data, with a time lag of around six to seven weeks. From this month, the publication will now include a ‘flash estimate’ of the number of employees paid through the tax system for the previous month, a lag of less than three weeks. This flash estimate is based on less actual data than the final estimates, but still covers a large proportion of the employee population. Meanwhile, the new BICS and the new LMS will reduce the gap between gathering the data and publishing the results compared with existing surveys. The Department for Work and Pensions is also releasing some early management information relating to Universal Credit declarations. These faster outputs will help meet the need for timelier information at this challenging time.
Getting our methods right
The COVID-19 pandemic has had an impact on how we collect data. For household surveys, we have stopped face-to-face interviews and moved to using exclusively telephone data collection. For business surveys, while nearly all of them are now online, because a number of businesses have temporarily ceased trading it is more difficult to get responses.
For both types of survey these changes are likely to lead to a reduced response rate. In normal circumstances, we would impute a figure based on previous responses to the survey. However, in the current situation, past information is unlikely to reflect what is currently happening. We are therefore working with our specialists in methods to work out how we handle the likely reduction in response in a robust way. As the releases start to cover survey data from March, we will provide further details of how we are handling the impacts on our surveys.
Protecting our outputs
In common with so many other people, ONS staff have had to adapt to new working processes. In order to ensure we can keep abreast of publishing the key labour market datasets and free up resources to process some of the new data sources and to provide the additional analysis, we decided to suspend the publication (and, in one case, collection) of some of the less-used datasets. We published a fuller statement on this on 3 April 2020.
David Freeman is head of Labour Market and Households at the ONS.