Labour productivity by economic activity

The COVID-19 pandemic has profoundly disrupted the relationship between declared employment and actual work performed, particularly due to the widespread use of short-time working compensation (RHT). These distortions have weakened the usual measurement of labour productivity by economic activity, due to potential biases that may now exist between the value produced and the labour input used to generate it.

The project launched by the FSO aimed to correct the labour input used in growth and productivity statistics (WPS). This correction provides a measure by economic activity that is better aligned with the actual volume of work, while also enabling analyses that are more consistent with economic reality, particularly in times of crisis.

Three reports accompanied this process:

• Methodological note 2022 (Ley & Helbling): initial presentation of the experimental results for the reference year 2020 and the different correction variants considered, with the provisional method selected.
• Methodological Note 2024 (Nedyalkova & Assoulin): consolidation of all developments since 2022, further exploration of sources and formalisation of the final method.
• Final Conceptual Report 2025 (Helbling): summary of the work, clarification of the conceptual framework and preparation for the transfer of the new indicators to regular production.

The correction based on RHT and APG-COVID allowances made it possible to better capture the true extent of the decline in activity, particularly in the most affected sectors such as hospitality and culture. This approach corrected a significant overestimation of full-time equivalents and reflects a contraction in employment more consistent with the decline in gross value added. It revealed a very different picture of productivity trends at the height of the crisis in 2020. Without these adjustments to labour input, the productivity growth rate would have been negative (-1.8%), whereas in reality it was remarkably positive (+4.2%). The method is reproducible and sustainable. It is now applied to the series and will capture any new events that may impact productivity through a reduction in working hours.

Labour productivity by economic activity is no longer published under the experimental statistics label. After rigorous methodological validation, it is now integrated into regular production and published directly on the webpage dedicated to growth and productivity statistics (WPS) (page available in German and French only).

Following the coronavirus pandemic that broke out across the world in early 2020, a large majority of governments had to enforce several strict health measures. These severe measures, such as lockdowns or bans on group gatherings, resulted in the temporary closure or reduction of the hours worked of many businesses and activities.

The pandemic and health measures have had a very negative impact on the economy. Both production and consumption have fallen considerably in some economic activities. However, in order to avoid an economic disaster, especially in the labour market, governments have taken measures to mitigate the impact.

Companies have thus received unprecedented support in the form of short-time working compensation or loss of earnings allowances. This assistance enabled companies to maintain the employment contracts of people who were temporarily unable to work due to legally mandated closures or a drop in demand.

Macroeconomic indicators such as GDP should correctly reflect the effects of the pandemic and government measures on the economy. The same applies to the input used in the production process, without which there would be a risk of bias in the measurement of labour productivity. In terms of calculating labour productivity, the situation varies depending on the indicator chosen. If we are interested in hourly labour productivity, which is available for the total economy only, the data used take full account of the impact of COVID-19. However, for labour productivity by economic activity, the input used comes from the structural business statistics (STATENT). However, this source only partially reflects the pandemic's impact. This is because STATENT's calculation of employment in FTEs is based on contractual relationships involving the payment of OASI contributions. However, thanks to short-time working and loss of earnings allowances, a significant proportion of OASI contributions continued to be paid, even when employees were prohibited from carrying out their work. The data underlying STATENT cannot therefore fully reflect the declines in activity in some sectors. A specific series for the years affected by COVID-19 must therefore be constructed. 

The aim is to produce and disseminate a measure of labour productivity by economic activity that is consistent with the economic reality. To do this, it is essential that the labour input reflects the amount of work actually done. Therefore, new input, corrected for short-time working and loss of earnings allowances that also corresponds to the reference framework for macroeconomic analyses has been developed. This new statistical product complements the information in STATENT, which should continue to reflect the contractual relationship.