[Economy] Labour productivity in Europe — 10 questions about who produces most per worker and why (2000-2024)

What this dataset contains

eurostat_labour_productivity_nuts3 (DataCivicLab, source Eurostat) measures nominal labour productivity — value added per person employed — by NUTS2 region, for all EU countries, from 2000 to 2024.

Coverage: 122K rows, 15 columns, 2000–2024
Granularity: regional (NUTS2) and provincial (NUTS3), annual
Metrics: euro per person employed (EUR), euro per hour worked, employment

---

Questions the data can answer

1. Why does Luxembourg produce 3× more per worker than Bulgaria?

Luxembourg: ~€137,000 per employed person (2024)
Bulgaria (Severozapaden): ~€41,000

A worker in Luxembourg produces 3.3× the value of a worker in Bulgaria. Is this real productivity, or does it reflect industry mix (finance vs agriculture), price levels, or cross-border commuting?

2. Which EU regions have the highest productivity?

The top tier includes Luxembourg, Île de France (€121k), Bruxelles (€112k), and Hamburg (€103k). All capital/city regions with strong financial services and corporate HQs. What does this tell us about the geography of high-value work in Europe?

3. Are Eastern EU regions converging?

Many Eastern European regions have grown faster in percentage terms (Poland +80%, Romania +110% since 2000). But in absolute euro terms, the gap with Western Europe remains enormous. At this pace, will they ever catch up?

4. Is the German industrial engine slowing?

Stuttgart (€103k), Oberbayern (€100k), Hamburg (€103k) — Germany's top regions remain highly productive. But their growth rate since 2010 has been modest compared to Eastern peers. Post-2022 energy crisis impact?

5. Why does Ireland have such extreme regional variation?

Ireland's NUTS2 regions show massive differences: Southern (€176k) vs Border (€58k). Likely driven by multinational tax structuring. How much of this is real productivity vs transfer pricing?

6. Are the Nordic countries converging or diverging?

Oslo (€107k) and Stockholm (€92k) lead, but remain below Luxembourg or Paris. Finland and Denmark show less internal variation than Sweden — more homogeneous economies or better data?

7. Which regions grew fastest since 2000?

Eastern European capital regions: București-Ilfov, Warszawski stołeczny, Bratislavský kraj — all grew 150-200% in 24 years. But they started from a very low base. Can secondary cities (Kraków, Timișoara, Brno) replicate this?

8. Is Southern Europe permanently behind?

Mediterranean regions — Southern Italy, Greece, parts of Spain and Portugal — cluster in the lower-middle tier (€60-80k). They grow in line with the EU average but never close the gap. Structural or cyclical?

9. Does hourly productivity tell a different story?

This dataset also includes productivity per hour worked. Countries with shorter working hours (Germany, Netherlands, Nordic) tend to look better on hourly basis. Italy, with long working hours, looks worse per person than per hour. Which metric is fairer?

10. How much does capital region bias distort national averages?

In many countries, the capital region (Île de France, Madrid, Wien, București) produces 30-60% more than the national average. In others (Berlin vs Hamburg, Rome vs Milan), the gap is smaller. Does the capital region drive the rest of the economy, or just concentrate it?

---

How to explore

The dataset is available via MCP: eurostat_query(slug='eurostat_labour_productivity_nuts3', sql='...')

-- Top 20 most productive regions (2024)
SELECT geo_label_en, ROUND(value) AS prod_per_person
FROM data
WHERE unit='EUR' AND na_item='NLPR_PER' AND nuts_level='NUTS2' AND year=2024
ORDER BY value DESC LIMIT 20;

-- Regions with fastest growth since 2000
SELECT geo_label_en,
  ROUND(AVG(value) FILTER (WHERE year=2000)) AS y2000,
  ROUND(AVG(value) FILTER (WHERE year=2024)) AS y2024,
  ROUND((AVG(value) FILTER (WHERE year=2024) / NULLIF(AVG(value) FILTER (WHERE year=2000), 0) - 1) * 100) AS growth_pct
FROM data
WHERE unit='EUR' AND na_item='NLPR_PER' AND nuts_level='NUTS2' AND year IN (2000, 2024)
GROUP BY geo_label_en
HAVING AVG(value) FILTER (WHERE year=2000) > 0
ORDER BY growth_pct DESC LIMIT 10;

Note: Data is at current prices (nominal). Real productivity growth is lower once inflation is accounted for. The dataset also contains data per hour worked (NLPR_HW).

[Gabrymi93]

Answer to Q1: Luxembourg produces 7.8× more per worker than Bulgaria — here's the data

The gap is even wider than the 3× I estimated in the question. Let me correct with real numbers:

Productivity per employed person (EUR, 2024, selected regions)

Region	EUR/worker	vs Severen tsentralen (BG)
Luxembourg (LU)	167,000	7.8×
Stockholm (SE11)	127,600	5.9×
Île de France (FR10)	123,200	5.7×
Hamburg (DE60)	119,700	5.6×
Lombardia (ITC4)	99,000	4.6×
Yugozapaden (BG)	39,800	1.9×
Severen tsentralen (BG)	21,500	1× (baseline)

Source: eurostat_query(slug='eurostat_labour_productivity_nuts3', sql="...")

What drives the gap?

1. Industry mix: Luxembourg's economy is dominated by financial services and EU institutions — high-value sectors. Bulgaria has more agriculture, manufacturing, and lower-value services.

2. Capital intensity: Luxembourg, Stockholm, and Paris have massive concentrations of corporate HQs, finance, and tech — sectors with high capital-to-labour ratios.

3. Commuting effect: Luxembourg's productivity figure is inflated by cross-border workers (100K+ daily from France, Belgium, Germany) who are counted as employed but not as residents. The value they produce is attributed to Luxembourg's economy.

4. Price levels: The data is nominal — not adjusted for purchasing power. A €100K salary in Luxembourg buys less than €40K in Bulgaria.

5. Within-country variation: Bulgaria itself has a 1.9× gap between its best region (Yugozapaden, which includes Sofia) and its worst (Severen tsentralen). The capital region effect is universal.

Would anyone like to dig deeper into a specific aspect — commuting effect, sector mix, or PPP adjustment?

[Gabrymi93]

Coverage note on Q9 — the hourly productivity series (NLPR_HW) is not in the published dataset

Quick data note before anyone tries Q9: the discussion body says "the dataset also contains data per hour worked (NLPR_HW)". It doesn't. The published clean parquet for eurostat_labour_productivity_nuts3 contains only NLPR_PER (nominal productivity per person employed), with units EUR, NAC and PC_EU27_2020_MEUR_CP — no NLPR_HW, and no hourly productivity series anywhere else in the Lab (checked the eurostat_economy composite too, which exposes productivity per person only).

Q9 as written ("Does hourly productivity tell a different story?") is therefore not answerable with the current data. The hourly comparison would require a new pull of the upstream dataflow — NAMA_10R_3NLP covers both per-person and per-hour measures, so this looks like a pull-coverage gap rather than a source limitation. Happy to file it as a candidate intake if the hourly angle is worth pursuing (it is: the per-person vs per-hour gap is exactly where Italy's long working hours show up).

The remaining questions (Q2-Q8, Q10) are all answerable with NLPR_PER as published.

[Gabrymi93]

Answer to Q3: yes — the East is converging, faster than the question assumed. But "enormous" is still the right word for the absolute gap.

On eurostat_labour_productivity_nuts3 (nominal EUR per person employed, NUTS2, unweighted averages, full series 2000-2024), comparing six Eastern countries (PL, CZ, SK, HU, RO, BG) with six Western ones (DE, FR, IT, ES, AT, NL):

Group	2000	2024	Growth
East6	10,276 €	44,196 €	×4.3
West6	59,747 €	102,021 €	×1.71

The West/East ratio collapsed from 5.8× to 2.3× in 24 years: the East went from 17% to 43% of the Western level. The claim in the question is understated: Poland grew +265% (not +80%) and Romania +881% (not +110%) — the full country ranking is RO +881%, BG +504%, LT +491%, EE +450%, LV +387%, SK +380%, CZ +313%, PL +265%, HU +234%, against DK +11%, EL +34%, IT +48%, FR +62%, DE +78%.

But the absolute gap is still enormous. The 2024 top tier is Ireland (Eastern & Midland 233.9k, Southern 206.7k — tax-structuring inflated, see Q5), Luxembourg 167k, Brussels 148.2k. The bottom tier is entirely Eastern/Southern: Šumadija and Western Serbia 15-19k, Bulgarian regions 21-28k, Romania's Nord-Est 23.7k, Epirus 33.5k. The most productive region produces ~15× the least — the gap in euro terms is as wide as the growth gap was wide in reverse.

Will they ever catch up? At the 2000-2024 rates the remaining 2.3× gap would close in roughly three decades — but that extrapolation is fragile: convergence rates are not constant, and the East's rapid phase was powered by EU accession (2004-2007) and a low starting base, both one-off factors. The honest answer is: relative convergence is real and fast, absolute catch-up is not imminent.

Two caveats: these are nominal values — part of the East's growth is inflation and price-level convergence, so the real gain is smaller than ×4.3; and the averages are unweighted across regions, so the East's figure leans on capital regions (București, Bratislava, Warszawa) and overstates the average regional worker.

Want to explore Q8 next — is Southern Europe (IT, EL, ES, PT) permanently behind, or does it show the same convergence pattern the East does?

Query (DuckDB, public clean parquet):

SELECT grp,
       ROUND(AVG(CASE WHEN year = 2000 THEN value END), 0) AS y2000,
       ROUND(AVG(CASE WHEN year = 2024 THEN value END), 0) AS y2024
FROM (
  SELECT geo, year, value,
         CASE WHEN LEFT(geo, 2) IN ('PL','CZ','SK','HU','RO','BG') THEN 'East6'
              WHEN LEFT(geo, 2) IN ('DE','FR','IT','ES','AT','NL') THEN 'West6'
         END AS grp
  FROM data
  WHERE unit = 'EUR' AND na_item = 'NLPR_PER' AND nuts_level = 'NUTS2'
    AND year IN (2000, 2024) AND value IS NOT NULL
) t
WHERE grp IS NOT NULL
GROUP BY grp;