The Two Europes: Why the Map of Decline Is Upside Down

In Part 1 I took apart the chart everyone shares — the one where Europe slides from a quarter of the world economy to a fifteenth. The numbers on it were wrong; the trend was real; and the honest scoreboard, once every figure was fixed, still showed a Europe that was smaller than it used to be, behind the United States on output per person, and ahead of or behind China depending on which ruler you picked up.

But every chart in that piece — the corrected ones included — committed the same quiet error. It treated "Europe" as a single line. One number, one trajectory, one fate.

It isn't one economy. It's at least two, moving in opposite directions, and when you average them into a single falling line you destroy the most important fact on the continent.

The continent that stopped being one number

Here is the move that the world-versus-Europe chart can't make, because it only has room for one European line. Stop comparing Europe to the world. Compare Europe to itself.

The cleanest way to do that is to stop using dollars and world-GDP shares — which bounce around with the euro–dollar exchange rate and with how fast China is growing this quarter — and use a ruler that's internal to Europe. Eurostat publishes exactly that: each country's GDP per capita measured in purchasing power standards, indexed so that the EU average is exactly 100. Above 100 means richer than the European average; below means poorer. It strips out the currency noise and the price-level differences, and it asks one honest question: relative to your neighbours, are you gaining ground or losing it?

Then you do the thing the doom-chart never does. You don't look at the level in a single year. You look at the change — where each country sat in 2004, when the great eastern enlargement began, and where it sits in 2024, twenty years later. That difference, country by country, is the real map of Europe. And it looks nothing like the one everyone argues about.

The map is upside down

Look at the continent the way the data actually draws it.

Green is a country that climbed toward the European average over those twenty years. Red is one that slipped away from it. And the pattern is unmistakable, almost embarrassingly so: the green is in the East, and the red is in the West and the South.

Romania went from 35 to 78 — from barely a third of the European average to better than three-quarters of it, in a single generation. Poland went from 51 to 80. Lithuania climbed from 53 to 89, Bulgaria from 35 to 66, the rest of the Baltics and the Visegrád countries close behind. These aren't rounding errors. They are the fastest sustained convergence of living standards anywhere in the developed world this century, and they happened on the supposedly declining continent.

Now find the red. France slid from 111 to 99 — from a tenth above the European average to just below it. Italy fell from 108 to 97. Spain dropped from 101 to 92. Austria gave up sixteen points, Finland and Sweden each gave up seven. And Greece — Greece collapsed from 97 to 70, a twenty-seven-point fall, the steepest on the continent. The countries that founded the European project, the ones the doom-chart imagines when it says "Europe," are the ones losing ground inside it.

Rank the moves and the divide is even starker.

Every single one of the biggest gainers joined the Union in 2004 or later, from the East. Every single one of the fallers is in the old West or the Mediterranean South. There is no overlap, no messy middle. The line between the Europe that's rising and the Europe that's sinking is almost exactly the old Iron Curtain — running the other way.

Which ruler, again

Now I have to do the thing that the doom-charts never do, which is argue against my own dramatic map. Because there's a trap in that index, and it's the same trap that makes the world-versus-Europe chart misleading: a relative number can move because you changed, or because everyone else did.

The index is share of the European average. It's a zero-sum ruler by construction — if the East gains twenty points against the mean, someone has to lose them. So some of the "red" in the West is not collapse. It's arithmetic. When Poland and Romania surge toward the average, they pull the average up, and a France standing still gets reclassified from "comfortably above" to "just below" without a single French factory closing. Relative decline and absolute decline are different diseases, and an honest map has to say which one it's looking at.

So let me say it. In absolute terms — real output per person, each country measured against its own past — most of Western Europe did not shrink. It grew, slowly, and was overtaken. That's a real problem, but it's the problem of a runner who jogs while others sprint, not one who falls down.

Two countries are the exception, and they're the ones that should keep European policymakers awake.

That distinction is the whole discipline of reading these numbers. The East is a genuine, absolute boom — Poland's economy grew about 3.7 percent a year for two decades, Romania's about 3 percent, real and compounding. The Western "decline" is mostly relative, the slow vertigo of being passed. And the Southern stall — Italy, Greece, parts of Spain — is the one place where the relative and the absolute agree, and the picture is genuinely bad. Three different stories, and the single falling "Europe" line erases all three.

The biggest quiet success in modern Europe

Pull the camera back to twenty years of motion and you can watch the two Europes move.

That widening, then closing, gap is the most important thing the doom-charts never show you. When the eastern members joined, their average sat below half of the European mean. Today it's near 80 percent and still climbing. Central and Eastern Europe as a whole went from 43 percent of the EU average in 2004 to 69 percent by 2023. That is the single largest peaceful convergence of incomes in modern European history — hundreds of millions of people hauled from the post-communist wreckage to within sight of Western living standards, inside one political and legal union, without a war or a conquest. If a development economist had predicted it in 1995 they'd have been laughed out of the room.

And underneath it, the southern line sags. The scissor isn't just the East rising. It's the East rising while the South sinks toward it — Italy and Greece drifting down to meet Poland and the Baltics drifting up. Sometime in the next decade, on current trajectories, Warsaw passes Lisbon and Athens for good, if it hasn't already. The map of "rich Europe" and "poor Europe" that every Western European over forty carries in their head is quietly being redrawn, and almost nobody has updated the picture.

This is why "is Europe declining?" is the wrong question. Europe is simultaneously running the most successful catch-up program on the planet and the most worrying stagnation in the developed world, in different postal codes, and the aggregate tells you neither.

Why the factories went east, and the south got stuck

None of this is mysterious once you stop looking at the scoreboard and start looking at the machinery.

The East converged because the single market did exactly what it was designed to do. Capital, supply chains and know-how flowed to where labour was capable and cheap. German carmakers built their second and third factories in Czechia, Slovakia, Poland, Hungary — not as charity but as cold logic, and the wages and skills followed the plants. Structural funds paid for the roads and the grids. A generation of engineers trained to Western standards and stayed home because home now had the jobs. Convergence wasn't a miracle; it was gravity, finally switched on by membership.

The channels are concrete, not abstract. Slovakia became the largest per-capita car producer on earth; the Visegrád four together turned into the assembly belt of the German auto industry. Estonia did it differently — it built a digital state and a startup density per head that the West still studies, and a lineage of companies from Skype to Wise to Bolt came out of a country of one and a third million people. Romania grew an IT and engineering export sector that now staffs the back ends of Western firms who've never set foot in Cluj. And then the 2020s handed the region a gift: the pandemic broke the case for ten-thousand-kilometre supply chains, the war next door broke the case for depending on Russia, and a wave of nearshoring started pulling production out of Asia and back toward Łódź and Timișoara. The East didn't just catch the old industrial wave. It's catching the next one, the one that prizes a factory you can drive to and a legal system you can trust.

The South got stuck for the mirror-image reason. It had the single market without the safety valve. For a country like Italy, the euro removed the one tool a stagnating economy has always used to claw back competitiveness — a weaker currency. You can't devalue the euro from Rome. So the adjustment that would once have happened in the exchange rate had to happen in wages, jobs and a decade of grinding internal deflation instead, and politically that adjustment mostly didn't happen at all. Italy spent twenty years not-quite-reforming and not-quite-devaluing, and the flat line on the chart is the result.

And the West — the comfortable core — has a third problem, the one Mario Draghi's competitiveness report was honest enough to name. Strip out information and communications technology and European productivity roughly matched the United States across 2000–2019. Europe didn't fall behind at everything. It fell behind at the one thing that compounds: digital technology. Four of the world's top fifty tech firms are European — four, not forty. Around 70 percent of the world's foundational AI models have been built in the US since 2017. The core didn't collapse; it missed the single most important industry of the era, and it's now paying the interest. (Part 1 has the long arc of that productivity miss in a single chart — output per hour against the US, near-parity in the 1990s, back below 80% today.)

The catch in the catch-up

Before I get triumphant about the East, the same discipline that made me suspicious of the doom-chart makes me suspicious of the boom. There are two catches in the catch-up, and an honest map has to show both.

The first is hiding in the denominator. The metric is GDP per capita — output divided by people — and one fast way to raise a ratio is to shrink what's underneath it. The East didn't just produce more; it exported people. After accession, the borders opened and the young and ambitious went west to earn western wages. Romania, Bulgaria, the Baltics and Croatia have lost extraordinary shares of their populations since 2007 — Latvia, Lithuania and Bulgaria are among the fastest-shrinking countries on the planet, having given up a fifth or more of their people in a generation. Some of the dazzling per-capita convergence is real, hard-won productivity. Some of it is simply that the capita emptied out. A village where the productive young have left for Munich will post a rising output-per-remaining-resident and call it progress, and the index can't tell the difference. The convergence is real; it is also, in part, a measurement of who's left.

The second catch is the ceiling, and it's the same wall Italy hit from the other side. Climbing from a third of the European average to three-quarters by assembling other people's designs is the easy stretch. The last quarter — closing the final gap and holding it — requires the thing the West has and the East mostly imported: home-grown innovation, owned intellectual property, brands and frontier research that capture the margin instead of renting it out. And the cost advantage that pulled the factories east is already eroding as eastern wages rise toward the West's. This is the textbook middle-income trap, and it would be a cruel irony if the East spent two decades sprinting away from the post-communist starting line only to slam into the exact stagnation that's pinning the South — assembly without ownership, growth without frontier.

That second catch is also where my own work lives, because the timing is brutal. The cheap-skilled-assembly-platform model the East rode up the chart is precisely the model that AI and automation are about to make cheap everywhere. The comparative advantage of "capable engineers at a discount, running a process designed elsewhere" is exactly the advantage that agentic systems erode first. Which means the East cannot treat the cognition stack as a luxury to buy later. For them, owning a piece of it is the difference between breaking through the ceiling and decorating it.

What a builder sees

Here's where I stop being a tourist with a spreadsheet and tell you why I actually care.

If you build things in Europe, the two-Europes map is not a melancholy curiosity. It's a directional signal about where the continent's growth, hunger and talent now live — and increasingly, that's the East. The engineers I work with in Warsaw, Bucharest, Tallinn, Kraków are not in a declining economy and they know it. They've watched their countries quadruple their relative position in their own lifetimes. They are not nostalgic, not defensive, not waiting for permission. That psychology — the lived experience of things getting dramatically better — is worth more to a hard technology project than any subsidy, and it has migrated east while the Western conversation curdled into managed decline.

But the convergence has a ceiling that the factories-and-roads model can't break through, and it's the same ceiling the whole continent is hitting. You can converge a long way by importing other people's supply chains and assembling other people's designs. You cannot converge the last stretch — or hold the gains — without owning the technology that sets productivity from here on. And that technology is now a stack: compute, models, energy, and the question of who controls them.

This is what "strategic autonomy" should mean once you scrape the Brussels varnish off it. Not protectionism, not nostalgia for national champions. The plain difference between setting prices and taking them in the one industry that will write the next twenty years of every line on every one of these charts. Draghi priced the reset at €750–800 billion a year — roughly four and a half percent of EU output, a scale of effort not seen since post-war reconstruction. Enormous. Also, set against the alternative of a second lost Italian generation spreading west, almost cheap. The continent is sitting on the private savings to fund it and arguing about the savings instead of deploying them.

The two-Europes map reframes that whole argument. The pessimist looks at the falling Western line and sees a continent in decline. I look at the rising Eastern line and see a continent that has already proven, twice now — first in the post-war West, then in the post-communist East — that it knows exactly how to converge fast when it points its capital and its institutions at a clear target. The capacity is demonstrated. What's missing is the target, and the will to fund it.

Retire the one-line chart

So I'd retire the "Europe versus the world" chart, not because its trend is fake — the slice really did shrink — but because it answers a question that was never the interesting one. Whether Europe is 16 percent or 18 percent of world GDP is a snapshot, and it's downstream of decisions made years earlier in domains a four-line chart can't see.

The map that matters is the internal one. It says Europe is not a single fading economy but two diverging ones — a converging East that is the quiet success story of the age, and a stalling West-and-South that includes a couple of genuine, frightening declines. It says the obituaries are being written for the wrong continent and pointed at the wrong half. And it says the lever that could pull both halves forward together is the same one I spend my days on: who owns the means of cognition, and whether Europe decides to own it here.

The chart on the screen is the weather report. The convergence machine underneath it — the one that already moved a third of a continent most of the way to the European average — is the climate. Stop arguing about the weather. Go build in the climate. And if you're choosing where, look east, where they still remember that the line is supposed to go up.

Sources and notes

• The core metric is GDP per capita in purchasing power standards, indexed to the EU average (EU = 100), from Eurostat (series tec00114). 2004 values are the published figures for that year; 2024 values are Eurostat's preliminary purchasing-power-parity estimates published in March 2025. Figures are rounded to whole index points; the country tiles and the ranking show the change in that index between the two years, in percentage points.
• The convergence headline — Central and Eastern Europe rising from about 43 percent of the EU average in 2004 to about 69 percent by 2023 — is from European Commission and Bruegel analyses of east–west convergence. The bloc-average "scissor" lines use those pinned endpoints; intermediate years (2009, 2014, 2019) are interpolated along the documented trajectory and are indicative, not exact annual datapoints.
• Italy's lost two decades (real GDP per capita essentially flat since 2000 — uniquely among large advanced economies) and Greece's contraction (real GDP development averaging about −0.3 percent a year across 2005–2024) are from IMF World Economic Outlook and Eurostat data; Poland's (~3.7 percent) and Romania's (~3.0 percent) average annual real growth over the same span are from the same sources.
• Ireland and Luxembourg are flagged and excluded from the ranking. Ireland's index is inflated by multinational profit-shifting (its GDP overstates domestic income badly); Luxembourg's is inflated by cross-border commuters who add to output but not to the resident population in the denominator. Neither number reflects living standards, and including them would distort the picture the essay is about.
• The competitiveness figures — European productivity matching the US outside ICT across 2000–2019, four of the world's top fifty tech firms being European, ~70 percent of foundational AI models built in the US since 2017, and the €750–800 billion annual investment estimate — are from Mario Draghi, The future of European competitiveness (European Commission, September 2024). The cloud-concentration point (three US hyperscalers serving the majority of the European market) is from the same report.
• The world-GDP-share framing (the EU falling from roughly a quarter of world output around 1990 to about 16 percent today, the US holding near 26 percent) uses IMF and World Bank figures at market exchange rates; as always, that share also moves with the euro–dollar rate, which is part of why the internal, purchasing-power ruler used here is the more honest lens for comparing Europe to itself.
• One deliberate simplification: "East," "South" and "Core West" are convenient groupings, not hard categories. Slovenia and Czechia were already near the Western range at accession; Portugal sits culturally South but trends mildly positive; the Nordics are their own story. The blocs capture the dominant pattern, and the country-level map is there precisely so you can check the exceptions yourself.