Evaluating the Methods of Estimating Total Hours Actually Worked: Insights from Labor Market Statistics

This study evaluates methods for estimating total quarterly hours actually worked within enterprises, using known annual values obtained from a complete enumeration survey conducted by the Statistics Poland. Using data from Q1 2009 to Q4 2023, we compare the estimates across economic activity sections with official quarterly survey data from Statistics Poland. Our approach prioritises practicality, computational feasibility, and statistical integrity. The evaluated methods are classified using forecast accuracy metrics and taxonomic tools based on the distance from an abstract ideal solution. The analysis demonstrates that methods employing average paid employment as an auxiliary series are more effective than approaches focused on movement preservation. Litterman’s method, which minimises the weighted residual sum of squares, exhibits the highest forecast accuracy and the greatest resilience to external shocks such as the COVID-19 pandemic and the global energy crisis. Our findings provide useful insights for selecting optimal interpolation methods in labour market statistics from a complete enumeration survey by the Statistics Poland.