By Ryan Boyle

We have commented lately on the challenges the Bureau of Labor Statistics has faced in collecting their data on employment and prices.  These problems are not unique to the United States.  The U.K. Office for National Statistics (ONS) faced a reckoning over the poor quality of its workforce survey.  The prolonged effort offers lessons for any nation considering changes to its labor data.

The Labour Force Survey (LFS) launched in 1973 to measure household employment and job seeking.  Over time, its scope grew to encompass 600 metrics surrounding earnings, health, social mobility and benefits, nationally and for 22 sub-regions.  Data collection, particularly for new panelists, was primarily performed through in-person interviews.  The pandemic halted these interviews and upended the LFS.  Already facing a low response rate of about 40% before shutdowns, the response rate fell dramatically to 14% in 2023. 

With so few responses, estimates of many LFS metrics were no longer reliable.  Trends were internally inconsistent and clashed with outside sources, such as the LFS indicating rising unemployment while all other measures of the job market showed a recovery.  The ONS suspended publication of important statistics like unemployment for seven months in 2023 while it worked to improve the survey.  Enhancements have continually fallen behind, with a complete changeover to a “Transformed” LFS dragging on to 2027.  In the meantime, compensating controls are employed, like using a three-month average unemployment rate to reduce monthly volatility.  Policymakers like the Bank of England have been left with less reliable data at a critical time to judge the economic cycle.

Every national statistical agency should pay attention.  First, costly face-to-face data collection should have been replaced long before COVID magnified its shortcomings.  For any other data collector considering enhancements, there is no time like the present; the LFS demonstrates the difficulty of making changes on the fly.  The ONS did the right thing by benchmarking against other government and private reports to identify problems in their results; contemporary data from the nation’s revenue agency helped observers bridge the gap while the LFS was remedied.  To their credit, the ONS has published regular reports on data quality and progress making enhancements.  Any other survey adjustments should meet a similar standard of transparency.

Throughout the process, LFS updates have upheld the importance of credible public statistics.  Changes have been slow, but the end result should be a more reliable product.  We hope any other modifications to government data around the world are informed by this experience.