Review Articles

Improving timeliness and accuracy of estimates from the UK labour force survey

D. J. Elliott ,

Office for National Statistics, Newport, United Kingdom

P. Zong

Office for National Statistics, Newport, United Kingdom

Pages 186-198 | Received 31 Jan. 2019, Accepted 01 Oct. 2019, Published online: 23 Oct. 2019,
  • Abstract
  • Full Article
  • References
  • Citations


Estimates of unemployment in the UK are based on data collected in the Labour Force Survey (LFS). The data is collected continuously and the survey design is structured in such a way as to provide quarterly estimates. These quarterly estimates are published each month as ‘rolling quarterly’ estimates. Currently the Office for National Statistics (ONS) publish rolling quarterly estimates, and these have been assessed to be of sufficient quality to be badged as ‘National Statistics’. ONS also publish monthly estimates of a selection of labour force variables, but these are designated ‘Experimental Statistics’ due to concerns over the quality of these data. Due to the sample design of the LFS, monthly estimates of change are volatile as there is no sample overlap. A state space model can be used to develop improved estimates of monthly change, accounting for aspects of the survey design. An additional source of information related to unemployment is administrative data on the number of people claiming unemployment related benefits. This data is more timely than survey data collected in the LFS and can be used to provide early estimates of monthly unemployment.


  1. Commandeur, J., & Koopman, S. (2007). An introduction to state space time series analysis. Oxford: OUP. [Google Scholar]
  2. Durbin, J., & Koopman, S. (2001). Time series analysis by state space methods. Oxford: Clarendon Press. [Google Scholar]
  3. Elliott, D., Zong, P., Greenaway, M., Lacey, A., & Dixon, G. (2015). A state space model for labour force survey estimates: Agreeing the target and dealing with wave specific bias. [Google Scholar]
  4. Harvey, A. (1990). Forecasting, structural time series models and the kalman filter. Cambridge: Cambridge University Press. [Crossref], [Google Scholar]
  5. Harvey, A., & Chung (2000). Producing monthly estimates of unemployment and employment according to the international labour office definition (with discussion). Journal of the Royal Statistical Society Series A160, 5–46. [Google Scholar]
  6. Helske, J. (2017). KFAS: Exponential family state space models in R. Journal of Statistical Software78(10), 1–39. doi: 10.18637/jss.v078.i10 [Crossref][Web of Science ®][Google Scholar]
  7. Hendry, D., Doornik, J. A., & Pretis, F. (2013, June). Step-indicator saturation. [Google Scholar]
  8. Hindrayanto, I., Aston, J. A., Koopman, S. J., & Ooms, M. (2013). Modelling trigonometric seasonal components for monthly economic time series. Applied Economics45(21), 3024–3034. Retrieved from [Taylor & Francis Online][Web of Science ®], [Google Scholar]
  9. ONS (2016a). Labour force survey user guide: Volume 1 - lfs background and methodology 2016. Retrieved from [Google Scholar]
  10. ONS (2016b). Uk labour market: Nov 2016. Retrieved from [Google Scholar]
  11. ONS (2017). Single month labour force survey estimates: Jan 2017. Retrieved from [Google Scholar]
  12. Pfeffermann, D. (1991). Estimation and seasonal adjustment of population means using data from repeated surveys. Journal of Business & Economic Statistics9, 163–175. [Taylor & Francis Online][Web of Science ®], [Google Scholar]
  13. Pfeffermann, D., Feder, M., & Signorelli, D. (1998). Estimation of autocorrelations of survey errors with application to trend estimation in small areas. Journal of Business & Economic Statistics16, 339–348. [Taylor & Francis Online][Web of Science ®], [Google Scholar]
  14. Pfeffermann, D., & Tiller, R. (2005, November). Bootstrap approximation to prediction mse for state-space models with estimated parameters. Journal of Time Series Analysis26, 893–916. doi: 10.1111/j.1467-9892.2005.00448.x [Crossref][Web of Science ®], [Google Scholar]
  15. R Core Team (2016). R: A language and environment for statistical computing [Computer software manual]. Vienna, Austria. Retrieved from [Google Scholar]
  16. Silva, D., & Smith, T. (2001). Modelling compositional time series from repeated surveys. Survey Methodology27(2), 205–215. Retrieved from [Google Scholar]
  17. Skinner, H., & Holmes, D. (2011). Variance estimation for Labour Force Survey estimates of level and change: GSS Methodology Series 21. Retrieved from [Google Scholar]
  18. Steele, D. (1997). Producing monthly estimates of unemployment and employment according to the international labour office definition (with discussion). Journal of the Royal Statistical Society Series A160, 5–46. doi: 10.1111/1467-985X.00044 [Crossref], [Google Scholar]
  19. Tiller, R. (1992). Time series modeling of sample survey data from the US current population survey. Journal of Official Statistics8(2), 149–166. [Google Scholar]
  20. Time Series Research Staff (2017). X-13arima-seats reference manual [Computer software manual]. Washington DC, US. Retrieved from [Google Scholar]
  21. van den Brakel, J. A., & Krieg, S. (2009, February). Estimation of the monthly unemployment rate through structural time series modelling in a rotating panel design. Survey Methodology35, 177–190. [Web of Science ®], [Google Scholar]
  22. van den Brakel, J. A., & Krieg, S. (2015, December). Dealing with small sample sizes, rotation group bias and discontinuities in a rotating panel design. Survey Methodology,41, 267–296. [Web of Science ®], [Google Scholar]
  23. van den Brakel, J. A., & Krieg, S. (2016). Small area estimation with state space common factor models for rotating panels. Journal of the Royal Statistical Society: Series A (Statistics in Society)179(3), 763–791. Retrieved from [Crossref][Web of Science ®], [Google Scholar]
  24. Werner, B. (2006). Reflections on fifteen years of change in using the labour force survey. Retrieved from–discontinued-/volume-114–no–8/reflections-of-fifteen-years-of-change-in-using-the-labour-force-survey.pdf?format=hi-vis [Google Scholar]