Technology Diffusion through Trade within the EU

Report Structure

Five parts · 27 EU member states · 1995–2025 · OECD TiVA bilateral trade data matched with ILO labour productivity statistics

🔄

Part 01

How Trade Spreads Technology

Four channels through which goods, services and capital carry know-how

Imports of capital goods, learning-by-exporting, foreign direct investment, and knowledge spillovers. Two ideas frame the analysis — the "advantage of being behind" (Acemoglu, Akcigit & Johnson 2026) and "absorptive capacity" (Atkinson & Stiglitz; Rodrik).

📊

Part 02

Three Speeds of EU Productivity

A fractured post-2019 productivity landscape

EU frontier slowed in step with the world frontier (0.6% vs 0.5%). EU non-frontier slowed to half that rate (0.3%). ECE kept growing faster — but unevenly. A new ceiling appears at ~50% of frontier productivity.

🔀

Part 03

How EU Trade Reorganised After 2019

Quiet structural shift in geography and composition

More intra-EU sourcing, more middle-tier partners, less exposure to the world's most productive economies. Trade with high-productivity peers collapsed by 10–20pp across all groups. Trade polarises toward middle and ultra-high tiers.

🇪🇺

Part 04

The Four Central & Eastern European Economies

Romania, Poland, Hungary, Czechia — diverging trajectories

Two emerging digital models (Romania, Poland) and two assembly-anchored models (Czechia, Hungary). Romania and Poland punch above their weight in IT exports to non-EU markets. Hungary's FDI-driven assembly model appears to have exhausted its convergence potential.

🎯

Part 05

What It All Means

Four policy priorities to revive technology diffusion

Build absorptive capacity (not just connectivity); defend exposure to high-productivity partners; support the emerging digital corridor (Romania & Poland); diversify away from assembly-only models before the catch-up boost runs out.

Twelve Main Findings

Grouped into four themes — productivity slowdown, the 50% threshold, trade restructuring, and ECE divergence

📉

Theme 01 · Productivity Slowdown

EU productivity fractured into three distinct speeds

Findings 01 – 03

1Finding

▼

22%

World frontier lead over EU frontier (1995–2019 avg)

EU and world frontiers grew in tandem; non-frontier fell behind

In 1995–2019, EU frontier, EU non-frontier and world frontier grew in tandem. After 2019 productivity growth halved globally; EU and world frontiers decelerated in tandem. EU non-frontier growth fell to half the frontier rate.

3Finding

≡

3 tiers

Frontier · Non-frontier · ECE (post-2019)

A three-tier productivity structure emerged post-2019

After overlapping shocks, EU productivity split into three distinct layers: EU frontier in line with world frontier (CAGR 0.6% vs 0.5%), EU non-frontier at half that rate (0.3%), and ECE growing significantly faster (Romania 2.7%, Poland 2.3%).

6Finding

⚠

0.3%

EU non-frontier CAGR — half the frontier rate

EU non-frontier is caught in a competitiveness trap

With average productivity at half the EU frontier, EU non-frontier countries grew at only half the frontier's rate post-2019. Attributed either to a productivity gap too narrow to sustain easy technology absorption, or to frontier technologies (AI and ICT) being inappropriate for their institutional and skill contexts.

🎯

Theme 02 · The Backwardness Threshold

A ceiling appears at ~50% of frontier productivity

Findings 02 · 04 · 05

2Finding

↗

~50%

Threshold of frontier productivity

ECE confirmed the advantage-of-backwardness — but a threshold has appeared

In 1995–2019 ECE countries with the largest productivity gaps grew fastest. After 2019 ECE countries give evidence of a productivity threshold at approximately 50% of frontier productivity, beyond which the advantage diminishes — suggesting either a shrinking technology gap or an inability to absorb frontier technologies, particularly in AI and ICT.

4Finding

↻

RO · PL · HU

Confirmed Acemoglu et al. (2026) prediction

The advantage of technological backwardness holds — but only up to a threshold

Romania, Poland, and Hungary confirmed the prediction: the larger the productivity gap relative to the frontier, the faster the growth. However, this advantage appears to vanish at around 50% of EU frontier productivity, beyond which catching-up stalls or reverses.

5Finding

⌐

38%

Czechia's productivity vs EU frontier · CAGR 0.6%

Czechia may have crossed the backwardness threshold

At 38% of EU frontier productivity, Czechia's growth rate converged to the frontier rate (0.6% CAGR 2019–2025) — suggesting it has entered a zone where the technology-diffusion benefit diminishes, despite not yet reaching frontier levels.

🔀

Theme 03 · Trade Restructuring

EU trade has been quietly reorganised since 2019

Findings 07 – 09

7Finding

⤴

54% → 60%

EU frontier's share of imports from within the EU

EU trade shifted structurally toward intermediates and intra-EU sourcing

The share of final goods in both exports and imports declined across all groups, reflecting deeper GVC embeddedness. Additionally, the EU frontier's share of imports from within the EU jumped from 54% to 60%, signalling a deliberate supply-chain regionalisation driven by post-pandemic de-risking.

8Finding

↓

Downgrade

Frontier shifting exports to lower-productivity OECD partners

The EU frontier is downgrading its OECD trade-partner quality

After 2019, EU frontier countries shifted exports away from high-productivity OECD partners toward lower-productivity ones, while maintaining ties only with the very top (world frontier). This trade-partner downgrading reduces exposure to the productivity spillovers historically associated with high-standard markets.

9Finding

⇄

−10 to −20pp

Collapse in trade with "high"-productivity EU partners

Within the EU, trade is polarising toward middle and ultra-high tiers

Across all country groups, the share of trade with "high"-productivity EU partners collapsed by ~10–20 percentage points, redistributed toward the "middle" tier and outliers (Luxembourg, Ireland). This reflects both value-chain restructuring and a comfort-zone dynamic where countries gravitate toward peers.

🌍

Theme 04 · ECE Divergence

Four ECE economies, two structurally different futures

Findings 10 – 12

10Finding

⤵

48–60%

Share of OECD imports from low-productivity partners

ECE is deepening intra-regional trade while sourcing cheap inputs from outside the EU

All four ECE countries reduced trade with low-productivity EU partners and expanded ties with middle-tier EU partners — each other and newer EU members — indicating an emerging CEE sub-regional economy. Simultaneously, OECD imports became increasingly concentrated in low-productivity partners (48–60% of OECD imports), reflecting cost-driven supply-chain strategies.

11Finding

↑

RO 9% · PL 6%

IT exports as share of rest-of-world exports (2022)

Romania and Poland are emerging as a digital services corridor within ECE

Both countries lead ECE in IT export shares, particularly to non-EU markets (Romania at 9%, Poland at 6% of ROW exports). This signals endogenous capability building beyond traditional assembly — their most promising long-run comparative advantage.

12Finding

⊠

27%

Hungary services exports — the lowest in ECE

Hungary is the most structurally vulnerable ECE economy

Its FDI-driven assembly model has exhausted its convergence potential: the productivity gap with OECD trading partners stalled after 2012, services exports remain the weakest of the four (27% in 2022), and IT export presence is minimal (3–4%). Without structural diversification, the model faces a ceiling with no clear escape route.

Productivity Growth — Before and After 2019

CAGR of labour productivity (USD per worker, ILO data) · Two periods compared

Country / Group	1995–2019 CAGR	2019–2025 CAGR	% of EU Frontier (2019–2025)
World frontier	1.2%	0.5%	122%
EU frontier	1.0%	0.6%	100%
EU non-frontier	1.3%	0.3%	~50%
Czechia	2.5%	0.6%	38%
Hungary	2.0%	1.0%	30%
Poland	3.3%	2.3%	31%
Romania	3.2%	2.7%	26%

The 50% Backwardness Threshold

A new ceiling appears at roughly half of EU frontier productivity

The 'advantage of being behind' works — but only up to ~50% of EU frontier productivity

Productivity growth (2019–2025) vs productivity level (% of EU frontier, 2019–2025 average)

Figure 5. The 50% backwardness threshold. Source: ILO statistic on labour productivity, 2025. Points sized for visibility; ECE economies sit well below the threshold and still grow fast; Czechia, the EU non-frontier and the EU frontier cluster around the world frontier growth line (0.5%).

Diffusion works · Below 50% of frontier

Romania & Poland · the fast-catch-up zone

Sitting at 26% and 31% of EU frontier productivity, Romania (2.7%) and Poland (2.3%) still convert imported technology into measurable productivity gains. The "advantage of being behind" still operates: there is enough to copy, and absorptive capacity is sufficient. This is the structural zone where EU funds, FDI and learning-by-exporting deliver the largest returns.

Diffusion stalls · Around 50% of frontier

EU non-frontier & Czechia · the ceiling zone

Czechia (38% of frontier) has slowed almost to frontier-level growth. The EU non-frontier — at roughly 50% of frontier — has slowed even more (0.3% CAGR). Once a country reaches ~50% of frontier productivity, the easy gains from copying frontier technologies start to vanish. From that point on, growth depends on the country's own innovation, not on imported know-how.

Four ECE Economies · Two Diverging Models

Romania, Poland, Czechia and Hungary entered the EU as similar post-communist economies. Twenty years on, they are following visibly different trajectories.

🇷🇴 Romania · Emerging digital

Hybrid model: industrial base + rising IT/services exports to global markets

Productivity (% of frontier)26%

CAGR 2019–20252.7%

IT exports to non-EU9%

Services share45%

Industry exports fell from 72% to 47%; services rose from 26% to 45%. IT exports to non-EU markets reach 9% — the highest of all four ECE economies. Romania's most promising long-run advantage is its IT and services trajectory to global markets.

🇵🇱 Poland · Regional hub

Most complex model: domestic market depth + IT/services corridor with Romania

Productivity (% of frontier)31%

CAGR 2019–20252.3%

IT exports to non-EU6%

Services share36%

The largest ECE economy by trade volume, with the fastest deindustrialisation: industry exports fell from 67% to 37% by 2022. IT exports reach 6% in non-EU markets. Poland combines assembly-hub characteristics with regional-hub ambitions and a strong domestic market.

🇨🇿 Czechia · Sophisticated assembly hub

Sophisticated assembly hub — but productivity may have hit the "backwardness threshold"

Productivity (% of frontier)38%

CAGR 2019–20250.6%

Industry share42%

OECD imports from low-tier60%

Industry exports at 42% (down from 73%); services around 42%. 60% of OECD imports come from low-productivity partners — the most aggressive cheap-input sourcing of all four. Czechia's productivity, at 38% of frontier, is now growing only as fast as the frontier itself — suggesting it has hit the backwardness threshold.

🇭🇺 Hungary · Assembly model under stress

FDI-driven assembly model appears to have exhausted its convergence potential

Productivity (% of frontier)30%

CAGR 2019–20251.0%

IT exports3–4%

Services exports27% (lowest)

Services exports just 27% — the lowest in ECE. IT exports 3–4% — the smallest digital footprint. Trade structure heavily locked into automotive and electronics FDI. The productivity gap with OECD trading partners stalled after 2012. The FDI-driven assembly model appears to have exhausted its convergence potential.

Four Policy Priorities to Revive Technology Diffusion

What it all means for EU industrial, trade and cohesion policy

Priority 01

Build Absorptive Capacity

Trade flows are necessary but not sufficient. The EU's diffusion engine works only when the receiving country can absorb new technology — R&D, skilled labour, broadband, modern management and functioning capital markets. EU funds should weight absorptive capacity at least as heavily as physical infrastructure.

Priority 02

Defend Exposure to High-Productivity Partners

Re-shoring and de-risking are sensible responses to geopolitical risk. But the strongest knowledge spillovers come from interaction with the world's most productive economies. Distinguish strategic dependencies (semiconductors, critical minerals) from ordinary trade — don't let de-risking become a blanket retreat from competitive markets.

Priority 03

Support the Emerging Digital Corridor

Romania and Poland's IT export performance is the most promising structural shift in ECE since EU accession. It is a chance to build a genuinely competitive European software and services hub — but it is fragile. Invest in education, broadband, grid investment and regulatory stability to retain skilled software workers.

Priority 04

Diversify Away From Assembly-Only Models

Hungary's stalled convergence is the clearest warning sign in the data. FDI-driven assembly delivers decades of growth — but only until productivity reaches roughly half of the frontier. After that, the host economy needs indigenous services, R&D and high-skill activities. The window opens before the catch-up boost runs out, not after.

The Bottom Line

Technology diffusion through trade is not broken — but it is bending in ways that quietly favour the EU's strongest economies and disadvantage its middle. The EU's productivity story over the next decade will be written less by trade volumes and more by who Europe trades with, and what its firms can do with what they import.

Technology Diffusion through Tradewithin the EU