Somewhere between Bratislava and Brno, genuine anxiety has settled over Central and Eastern Europe’s economic policymakers. The formula that powered thirty years of remarkable catch-up growth (cheap labour, EU structural funds, foreign factories) is losing its potency. What has emerged in its place is less a strategy than a predicament: economies too expensive to compete with lower-cost neighbours on the periphery, not yet innovative enough to run with the Nordics and the Dutch. The World Bank has a name for it. Welcome to the middle technology trap.
The term captures something the GDP growth figures have obscured. Poland, Croatia, Bulgaria, Romania have all converged on Western European living standards much faster than most economists predicted. Polish productivity grew at 3.3 per cent per year in the decade before 2008. From 2010 to 2019, that figure dropped to 1.9 per cent. In Czechia and Slovenia, wages are now converging faster than output per worker, a combination that tends to concentrate minds among finance ministers. The Bank’s 2026 EU Regular Economic Report, published this week, frames it dryly: the ‘convergence dividend’ is exhausting itself.
Geography isn’t helping. High-tech industries account for only 15 per cent of private R&D in Europe, against roughly 85 per cent in the United States. CEE countries cluster in the middle: automotive components, light manufacturing, business services that require more skill than a Chinese factory floor but less than a Finnish semiconductor lab. Call it mid-tech, which the report calculates accounts for roughly half of all EU business R&D. It is not a bad place to be. Until it is.
R&D? What R&D?
Though they are anything if not familiar, the numbers behind the innovation gap are bleak. Romania and Bulgaria have spent less than 0.8 per cent of GDP on research and development since 2016. The EU average is 2.2 per cent. Patent applications per capita across the four main CEE economies sit at less than four per cent of Germany’s level. Fewer than one in five Romanian tech start-ups use AI or big-data analytics in any meaningful way. None of this comes as a great shock to anyone who has spent time in Bucharest or Sofia. The shock is that the gap, after three decades of convergence, has not narrowed.
The World Bank is diplomatically guarded here, but the underlying data suggest something beyond mere under-investment. Romania has more STEM graduates than most EU member states. The limiting factor is not talent, it is the institutional environment that talent refuses to stay in. One in five highly educated Romanians and Bulgarians has emigrated.
Poland is sometimes held up as the counter-example, and not without reason. Fourteen unicorns. An ecosystem value that has risen 150 per cent since 2020. Warsaw’s start-up scene draws regular comparisons with Tel Aviv, though usually from people who have not visited either recently. But even Poland illustrates the trap rather than escaping it. Venture capital investment runs at roughly 0.05 per cent of GDP, below the EU average. Government grants still account for around 34 per cent of total start-up funding across the region, which would be fine if grants reliably went to the right companies, which they do not. In Bulgaria, 80 per cent of public innovation support comes through grants rather than equity or loans. Croatia’s figure is 88 per cent. Grants are excellent at preserving existing structures; they are poor at disrupting them.
The VC that does flow skews heavily towards pre-seed and seed rounds. Over 70 per cent of CEE funds operate at that end of the spectrum. Series A and beyond, where start-ups turn into actual businesses with revenue and payroll, remains thin. The report notes that 4CEE start-ups achieve comparable revenue milestones to their Western peers with 40 per cent less capital, which sounds like admirable frugality until you realise it mostly reflects the difficulty of raising more.
The comparison with Israel is instructive and underused. Israel’s innovation success rested on specific structural pillars: military R&D with clear commercial spin-offs, a venture capital industry deliberately seeded by the state through the Yozma programme in the 1990s, and an unusually tight network of elite technical talent concentrated in a small geography.
CEE has some of these ingredients but not the combination. Defence spending may eventually supply the missing piece, Poland is now allocating 4.2 per cent of GDP to defence, and the Baltic states are heading towards four to six per cent, but the translation of military procurement into civilian innovation reliably takes a decade or more. The pipeline exists. The patience may not.
Taking away the ladder
There is an AI subplot running through all of this, and it cuts both ways. AI accounted for 45 per cent of all CEE venture capital investment in the first quarter of 2024. Polish AI start-ups doubled their fundraising total in 2023. The interest is real, the momentum genuine, and the timing, for once, reasonably well-aligned with global trends. The report introduces something called the MCPAT-AI framework, mapping the competitive risks that AI poses sector by sector. The analysis, however, suggests that automation is already eroding the mid-skill manufacturing jobs that built CEE’s industrial base, and AI is beginning to replace the business-process outsourcing roles that have underpinned its services sector. The very industries that provided the ladder are having their rungs removed.
The regulatory picture complicates matters further. The World Bank flags a pointed irony: GDPR reduced venture capital investment in data-driven European start-ups by more than a third and cut the number of apps available on Google Play by 13 per cent. The EU AI Act is now being layered on top. Compliance costs fall disproportionately on smaller firms, which is, with perfect inconvenience, precisely the structural profile of CEE economies. The regulation was designed with big-tech incumbents in mind. Its burdens land mostly on the startups trying to compete with them.
This, arguably, is the EU’s persistent innovation paradox: a single market designed to create scale, combined with a regulatory architecture that makes scale harder to achieve. CEE is not uniquely exposed to this tension, but its firms have fewer resources to absorb compliance costs than their Western counterparts. The AI Act’s risk-tiering means that high-risk applications (medical, legal, financial) face the heaviest requirements. These happen to be sectors where CEE has genuine technical strengths. The timing is unfortunate.
Where are all the people?
Demographics do not offer much comfort either. Bulgaria’s working-age population is projected to shrink by 35 per cent by 2050, the steepest decline in the EU. Poland’s by 30 per cent. Romania lost 1.1 million people, roughly 5.7 per cent of its population, between its 2011 and 2021 censuses. CEE governments have shown limited appetite for the kind of large-scale labour migration that might offset the losses. The factories still needing workers are competing for the same shrinking pool as the tech firms desperately short of engineers. Building an innovation economy on a collapsing demographic base is, to put it gently, awkward.
The regional variation within countries sharpens the picture. In Romania, the gap in poverty rates between the richest and poorest regions runs to 26 percentage points, the widest in the EU. The innovation economy concentrates in Mazowieckie and Lower Silesia in Poland, in Sofia in Bulgaria (which accounts for 87 per cent of the country’s start-ups), in Bucharest. The rest is largely left behind. EU funds have softened this, not reversed it.
The report calculates that catching up with the EU digital average could lift labour productivity across the four main CEE economies by six to eight per cent. Actually deepening digital intensity (using technology rather than merely installing it) could yield 10–15 per cent. That would be transformative. What stands between the region and those gains is mostly institutional: the quality of public universities, the willingness of regulators to let new entrants disrupt incumbents, and the degree to which talented people conclude that staying home is worth the trade-off.
A generation of disruption
None of that is insuperable. Estonia made the journey from Soviet occupation to digital republic in roughly a generation, though from a smaller and more homogeneous starting point, with a government that happened to treat institutional reform as a national survival project. Croatia has just recorded the fastest improvement in innovation performance of any EU member state over seven years, a 19.4 percentage-point rise against the EU average, primarily through targeted EU fund deployment and a tech sector that has quietly built strengths in fintech and SaaS. Whether that momentum survives the usual political headwinds is another question.
Nevertheless, Croatia’s trajectory deserves more attention than it gets. The country still runs one of the highest tourism dependency ratios in Europe, around 11 per cent of GDP, and the long-term exposure to climate-driven shifts in Mediterranean tourism is a structural risk that barely registers in innovation rankings. Innovation metrics measure inputs and patents; they do not measure what happens when July becomes economically unusable. Croatia’s upgrade to Moderate Innovator status is real, but the foundation remains narrower than the headline suggests.
The window, the World Bank gently implies, is not indefinite. The mid-tech industries that sustained CEE growth for thirty years face a generation of disruption. The demographic pressures will not wait for policy to catch up. And the AI wave (which might, under the right conditions, allow the region to leapfrog some of the intermediate development stages that constrained earlier catch-up economies) is already moving faster than most governments in the region can track, let alone direct.
The IMF estimates that the EU could raise GDP by seven per cent simply by cutting internal barriers by 10 per cent. That arithmetic has been available for years. So has the prescription: more equity finance, less grant dependency, better universities, lighter regulatory burdens on small firms, genuine cross-border capital markets. The Draghi report said much the same last year, at greater length. The Letta report before it. The gap between diagnosis and action in Brussels tends to be capacious.
The middle lane looks fine until the traffic around you starts accelerating. For CEE, that acceleration is already underway, in Warsaw’s AI scene, in Sofia’s nascent defence-tech clusters, in Zagreb’s fintech firms raising rounds in euros and listing ambitions in Amsterdam. The question is whether a generation of policymakers built on the convergence model can move fast enough to build something new before the old model stops working.
History offers mixed odds.
Photo: Dreamstime.
