Is productivity up, down, more important than elsewhere? Four institutions have positioned themselves and all four have a correct analysis that focuses on a different aspect. And none of them are considering AI at the moment. (Photo: Shutterstock)

Is productivity up, down, more important than elsewhere? Four institutions have positioned themselves and all four have a correct analysis that focuses on a different aspect. And none of them are considering AI at the moment. (Photo: Shutterstock)

The BCL, CSL, Statec and the Chamber of Commerce have each delivered diverging readings of productivity in Luxembourg over a few days. They all share one blind spot: none of them mentions the impact of artificial intelligence, despite it being the focus of international analyses about future productivity gains.

While artificial intelligence has become, elsewhere in the world, the main lever hoped for to boost productivity, the Luxembourg debate continues to be based on classic springs: wage costs, value added, labour retention, sectoral effects and indexation. The three macroeconomic documents recently published - the Banque centrale du Luxembourg’s (BCL) opinion on the 2026 budget, the Statec's Note de conjoncture and the Chamber of employees’ EcoNews – as well as the Chamber of commerce’s Baromètre économique offer divergent diagnoses, but have one striking point in common: none of them refers to AI as a factor in transforming productivity. At a time when the major international institutions are already assessing its potential effect on growth, this absence raises questions.

In its opinion, the BCL analyses productivity in detail, in particular via unit labour costs, sectoral trends and the relationship between value added and employment. It points out that, between 2022 and 2024, the Luxembourg economy experienced a marked contraction in activity in trade, business services, finance and construction, while employment remained dynamic. This lag explains some of the slowdown in apparent productivity. The diagnosis focuses on well-identified cyclical and structural factors, but not on the potential impact of advanced digital technologies. AI does not feature in the analysis, even though, in other countries, it is seen as a possible driver of reallocation, automation and upmarket tasks.

A pre-digital reading grid

The Chamber of Employees takes an almost opposite approach. In its EcoNews, it challenges the idea of a structural weakening in productivity by basing itself on nominal productivity, which shows that Luxembourg remains one of the most productive countries in Europe, behind only Ireland. One hour's work will generate almost 114 euros of added value in 2024, compared with 77 euros in Belgium, 70 euros in Germany and 63 euros in France. For the CSL, Luxembourg's comparative advantage is even growing. Here again, the analysis focuses on levels, international comparisons and methodological conventions. But no mention is made of the potential gains linked to the spread of AI, or of the risks of a technological slowdown. The debate remains rooted in a pre-digital reading grid where productivity depends essentially on volumes of activity and work.

Statec takes an intermediate and more technical stance. Its study 7.3 confirms a fall in hourly productivity in the non-financial market sectors from 2020, of around 2% per year, compared with an increase of 1.1% between 2000 and 2019. The Institute identifies factors such as labour hoarding, fluctuations in activity, real wages and certain cyclical variables. It highlights a very strong correlation between real wages and productivity, without concluding on the direction of causality. AI is mentioned only once, and only marginally, as a methodological tool: the study notes that machine learning techniques could improve econometric specifications. But the impact of AI on production, work organisation or the diffusion of innovations is not addressed. This silence contrasts with the numerous international publications that specifically examine the role of advanced automation in the productivity divergences observed over the past decade.

An open question

The Chamber of Commerce, on the other hand, shifts the debate to labour costs. In the Economic Barometer for the second quarter of 2025, 59% of companies surveyed believe that the cost of labour is their main challenge, and 55% cite the lack of skilled labour. For Christel Chatelain, Director of Economic Affairs, "there is a decoupling between rising labour costs and productivity". Carlo ThelenCarlo Thelen, director general, attributes much of this tension to the automatic indexation mechanism, which he believes penalises exporting companies and weakens competitiveness. The debate remains focused on labour costs and the institutional environment, as if the rapid technological transformations affecting global productivity were absent from the Luxembourg landscape.

This general absence of explicit reflection on the impact of AI is all the more remarkable given that Luxembourg has, in theory, favourable characteristics for its spread: a high level of labour productivity, a high intensity of high value-added services, an internationalised workforce and an entrepreneurial fabric accustomed to financial and regulatory technologies. While AI offers new levers for automating routine tasks, increasing process efficiency, improving data analysis or strengthening compliance functions, these potential effects are not currently included in any of the institutional diagnoses.

The Luxembourg debate thus appears out of step with discussions in the United States, the United Kingdom and several European countries, where AI is now perceived as a key factor in total factor productivity and a future determinant of real wage growth. In Luxembourg, the debate remains structured around three traditional axes: productivity measurement, sectoral dynamics and labour costs. The major technological shock currently underway has not yet been taken into account.

For the time being, the four institutions converge on one point: Luxembourg’s productivity is both high in level and weakened in certain branches since 2020. They differ on what this means for competitiveness, wages and the economy in the long term. But they all leave open a key question: could AI, as elsewhere, become the driving force behind a new productivity cycle? Or will Luxembourg, because of its particular economic structure, remain relatively impervious to this transformation? As the technological revolution gathers pace, this question will undoubtedly become inescapable for the Luxembourg economy.