Coverage of yesterday’s labour market statistics showed the Westminster political/media nexus at its worst. Rather than focus on close to three-quarters of a million more people getting work, the Westminster ‘narrative’ for politicians and journalists was instead about thousands of Romanians and Bulgarians supposedly having left the country. Their basis for this essentially fictitious story was a less than 3% change in a 0.4% subset of the Labour Force Survey (LFS).
Not one of the politicians and journalists leaping to conclusions seems to have read even the two paragraph Executive Summary of the ‘Information paper’ published by the Office for National Statistics (ONS) on using the LFS. The first line states ‘The primary purpose of the LFS is “the prompt publication of key aggregate, whole economy, indicators, for the integrated assessment of labour market conditions”‘. Its purpose is manifestly not to conclude how many Bulgarians or Romanians are entering or leaving the country in a particular quarter.
The ONS Information paper Executive Summary continues “Output from the LFS is quarterly since 1992. Each quarter’s sample is made up of five waves at three-monthly intervals. The sample is made up of approximately 41,000 responding UK households per quarter. Respondents are interviews for five successive waves at three-monthly intervals and 20% of the sample is replaced every quarter”.
So, this survey cannot possibly tell us how many Romanians and Bulgarians started working in the UK during the first quarter of this year, since four-fifths of the UK households surveyed were recruited to the survey last year, before we removed restrictions on Romanians and Bulgarians working here.
Moreover, to try to produce accurate information about the approximately 30.43 million people recorded as working, the ONS aims to undertake 41,000 surveys, a 0.13% sample. That is fine for estimating that 722,000 more people are in work than a year before, since the 0.13% of that 722,000 they have sampled equates to around 1,000 survey responses. We must though recognise that we don’t quite know that 722,000 more people are in work, due to random sampling error, but we can be pretty certain that the increase was somewhere between 650,000 and 800,000.
It is not however fine to use this survey to say that 4,000 Romanians and Bulgarian have left the country since the margin of error will be many, many times larger than the number cited. We can see this if we multiply that 4,000 by the 0.13% sample size. We discover that the past day’s news cycle about a supposed exodus of Romanians and Bulgarians from the country has been based on the survey response of just five people.
I tried my best to puncture this absurd media/political bubble on Newsnight last night.
