The two Europes
Europe’s security shock is deepening its economic divide
(I coauthored this post with Olivier Kooi who is a postdoc at the LSE.)
While we have spent a decade complaining about Europe’s stagnation, a real positive convergence story has been unfolding at the eastern edge of the continent. Poland entered the millennium at 34% of US per capita GDP; by 2030, the IMF projects it will reach 67%. Romania rises from 27% to 60%. Lithuania moves from 29% to 69%. Bulgaria goes from 23% to 53%. These countries’ citizens are still significantly poorer than Americans. But they have closed a large part of the gap with the global frontier. Across Central and Eastern Europe, the EU has functioned as a powerful convergence machine for countries willing to use it.
Meanwhile, at the other end of the continent, the picture is almost the mirror image. Portugal slipped from 64% to 57%. Spain fell from 72% to 61%. The richer laggards are more striking: France dropped from 86% of the US level to 71%, and Italy fell from 93% to 68%.
This split explains why the recent debate about whether Europe is “really” stagnating has been so unsatisfying. Gabriel Zucman and others such as Krugman and Saez point to the convergence numbers and argue that European growth has been underestimated. Philippe Aghion and others insist that Europe has fallen badly behind on productivity and innovation. Both are right, but about different countries. The Saez-Zucman story is the eastern half of the chart: rapid catch-up from a low base, driven by capital deepening and institutional convergence after EU accession. The Aghion story is the western and southern half: rich countries that have not kept pace with the US frontier in the technologies that now drive growth. The EU is not one economy moving slowly. It is two economies moving in opposite directions, and the disagreement about which story is true is really a disagreement about which Europe you are looking at. This fracture already existed before Russia’s full-scale invasion of Ukraine in February 2022.
The hope in Europe after the Draghi report was that the Russian threat to the east would supply the political urgency the reform agendas in the “stuck” part of Europe lack.
The correlation created by the crisis is the wrong one. The Russian shock is mobilising the countries where the Draghi problem is not most severe, and leaving the countries where it is most severe largely untouched. War is reforming Europe’s frontier. It is not reforming Europe’s rear, and on present evidence, it will not.
Mancur Olson saw this coming. In The Rise and Decline of Nations (1982), he argued that long-stable societies accumulate “distributional coalitions” over time. Producer associations, trade unions, professions, public sector constituencies, regional lobbies, and business incumbents each learn to maximise the rents they receive from the current institutional setup. None of them is large enough to wreck the economy on its own, but acting together they make any reform that imposes concentrated losses almost impossible to pass. Olson’s conclusion was that settled societies rarely reform themselves voluntarily. They reform when something forces the calculation behind the existing bargain to change.
The canonical example in this literature is the Meiji Restoration. In 1853, Commodore Perry’s warships entered Uraga Bay, at the mouth of Edo Bay, and demanded Japan open its ports. Within fifteen years, the humiliation had toppled the shogunate. The Meiji elites concluded they had to industrialise or be colonised, and built the railways, schools, and tax system to do it.
Continental Europe is Olson’s textbook case. Its democracies, while younger than Britain and the United States, are old by global standards. It has gone through two decades of low growth, so there is little surplus to pay losers off, and interest groups are used to fights being zero-sum. Many of these groups are made up of old people, so there is even less reason to accept short-term pain for longer-term benefits. This is why the agenda set out in the Draghi report is so hard to deliver. Almost everyone agrees that Europe needs more innovation, more energy investment, greater defence capacity, and larger, more productive firms. Almost every reform that would deliver these things creates concentrated losers today, and is therefore blocked.
War can change the bargain. Tim Besley and Torsten Persson, in their work on state capacity, have argued that an external threat to survival can function as a common-interest public good. When the existence of the state itself is at stake, elites become less willing to tolerate narrow rent-seeking and more willing to fund the things that hold the state together: tax administration, courts, infrastructure, schools, a standing army. The motivation is not enlightenment but the alternative, which is being conquered.
Besley and Persson made explicit when this mechanism works and when it does not. It works where common-interest institutions are strong enough that the state can credibly commit shared resources to common purposes. Take South Korea. When Park Chung-hee took power in 1961, North Korea looked economically and militarily stronger and had Soviet and Chinese backing; Park redirected an already-functioning state apparatus toward extraction, export discipline, infrastructure, and education. Israel and Taiwan followed similar paths, building capable states that spend heavily on defence while prioritising the growth that sustains long-term deterrence. The mechanism fails where the underlying institutions are weak: the threat then converts into repression instead, as in Pakistan or Egypt.
Europe is a third type that the existing theory does not quite accommodate. European states already have high common-interest capacity, but that capacity is pointed at consumption transfers rather than at growth-promoting public goods. A security shock in such a state will raise defence spending easily, because adding a new claim on the budget is what capable states do under threat. Whether it also redirects the fiscal stance toward growth depends on something harder: whether the shock is large enough to dislodge the pensioners and insiders who own the existing consumption budget. Defence is the easy first step; public investment and education are harder. We examine them below.
Part of Europe is mobilising
The empirical question we pose is whether the war accentuated the existing fracture. The countries closest to Russia were often either fast-converging economies or high-capacity Nordic states, while the countries most stuck in the Draghi sense were often farther away. So our question is: after 2022, did countries closer to Moscow reallocate more than their own previous path would suggest? The regressions in the appendix control for permanent country differences, common EU-wide shocks, and the separate year-by-year catch-up path of post-2004 entrants.
We find that, after controls, each 750 km (one standard deviation) closer to Moscow predicts about 0.30 percentage points of GDP larger increase in defence spending after 2022. The relationship is strong, as the left panel shows, and statistically significant (the standard error is 0.09).
The harder question is whether the war just led these countries to buy weapons or whether it also led them to invest in long-term growth. We also see increases in public investment for the countries that are close to Moscow. A state that is 750 km closer to Moscow would be expected to spend 0.39 percentage points of GDP more on government gross capital formation after 2022. Again the effect is statistically significant (standard error is 0.12). Finally, the evidence suggests that education spending also was less reduced in countries closer to Russia, by about 0.18 percentage points of GDP per standard deviation of proximity. The estimate is less precise than the defence result, but the gradient is in the right direction.
The combined bundle of defence, education, and investment moves by roughly 0.9 percentage points of GDP per standard deviation of proximity and is again significant. At minimum, frontier governments appear to be reallocating fiscal effort in a way that is broader than defence alone. Whether this becomes durable state-capacity building is too early to say.
Fiscal mobilisation is not only about adding new spending. It is also about whether existing claims on the budget crowd out future-oriented investment. In Europe, the most important of these claims is pensions. Pension spending moves slowly, and we see no post-2022 effect in the data. What the figure below shows is descriptive, not causal: countries far from Moscow already devote more of national output to pensions than countries close to it, and the gap is projected to widen.
According to our calculations from the European Commission’s 2024 Ageing Report, each additional 1,000 km from Moscow correlates with about 1.07 percentage points of GDP higher pension spending in 2022, and about 0.49 percentage points more projected growth by 2040 (Pearson correlations 0.28 and 0.32). The mechanism is demography and the maturity of pension systems, not Russian threat: eastern accession states have younger populations and pension regimes rebuilt later, on tighter terms. But the descriptive fact still matters for the larger argument. The rear of Europe carries the heavier consumption claim and feels the weaker security shock at the same time.
War may reform Europe’s frontier. It is much less clear that it can reform Europe’s rear. For the Draghi agenda, that is the central problem.
Fear of Putin will not do Draghi’s work
The Draghi report implicitly hopes that Europe’s two crises will reinforce each other. The competitiveness crisis must lead Europe to invest, innovate, build scale, and stop protecting incumbents. The security crisis is the reason Europe no longer has the luxury of drift. In the optimistic reading, the Russian threat supplies the political urgency that the competitiveness agenda has always lacked.
The evidence points to a more awkward conclusion. The security shock is creating political room for reform mainly where the competitiveness problem is not most severe. Poland, the Baltics, Finland, and Sweden are either closer to the administrative frontier or growing faster through convergence. The shock is not creating comparable room in the countries where the Draghi agenda is most needed. In Italy, Spain, and France, low productivity growth, ageing fiscal commitments, and protected incumbents are the heart of the problem, and the war is, for the median voter, still a foreign-policy event.
That is the missing piece in the European debate. Nobody has yet supplied a political mechanism that can make the rear of Europe invest, modernise, and grow. The countries that most need the Draghi reforms have to find a domestic reason to want them. That means a growth coalition strong enough to take on pensioners and other incumbents at the same time, which is something few European democracies have managed durably. Macron’s first term came closest, before the Gilets Jaunes ended that, as we discussed here. On present evidence, the mechanism will have to come from a fiscal crisis, or a larger external shock. Without it, Europe will not converge on a new growth model. At best, it will split: a mobilising frontier, and a comfortable rear drifting into managed decline.
Appendix: Data and regressions
The panel covers the EU-27 from 2000 to 2024. Defence and education spending come from Eurostat’s general-government expenditure by function dataset (gov_10a_exp), measured as a share of GDP. Public investment is general-government gross capital formation (ESA transaction P5), also as a share of GDP. Pension figures come from the European Commission’s 2024 Ageing Report. Distance is the great-circle distance from each capital to Moscow; proximity is its negative, standardised across countries. “Newcomers” are the thirteen states that joined in 2004 or later. “Post-2022” means the years 2022, 2023, and 2024. Since Russia’s full-scale invasion began in February 2022, including 2022 is natural but conservative: some 2022 budgets were already partly set before the shock.
The table reports two views of the same post-2022 shift. The first three numeric columns are raw country averages: the average change for the far tercile, the average change for the near tercile, and the unconditional near-minus-far difference. The final column reports the coefficient on standardised proximity to Moscow interacted with a post-2022 indicator from the fixed-effect regression. The regression includes country fixed effects, year fixed effects, and newcomer-by-year fixed effects. Country fixed effects strip out permanent differences between states. Year fixed effects absorb common EU-wide shocks. Newcomer-by-year fixed effects allow post-2004 entrants their own catch-up path in every year, so the coefficient does not simply pick up the East-West income gradient.
There are obvious limitations to this analysis. The post-2022 period is short, and outcomes such as education and public investment respond with long lags, so the education estimate may grow or fade as more data arrives. The fixed effects cannot rule out every shock that happens to correlate with distance from Russia after 2022.
References
Besley, Timothy, and Torsten Persson, 2008, “Wars and State Capacity,” Journal of the European Economic Association 6(2-3): 522-530.
Besley, Timothy, and Torsten Persson, 2009, “The Origins of State Capacity: Property Rights, Taxation, and Politics,” American Economic Review 99(4): 1218-1244.
Draghi, Mario, 2024, The Future of European Competitiveness, European Commission.
European Commission, 2024, The 2024 Ageing Report: Economic and Budgetary Projections for the EU Member States (2022-2070).
Olson, Mancur, 1982, The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities, New Haven: Yale University Press.
Extremely insightful, thank you, Luis. One comment: (1) your Figure 1 should be reflected in sovereign bond yields (apples-to-apples, i.e., yields on the euro-denominated bonds traded on luxse.com for the non-Eurozone countries). (2) In theory, these yields should create powerful incentives to "do something" or "stop doing something". (3) But the ECB distorts these yields by its yield support program—Cochrane, John H., Luis Garicano, and Klaus Masuch. Crisis Cycle: Challenges, Evolution, and Future of the Euro. Princeton, NJ: Princeton University Press, 2025. (4) So they do not reflect your Figure 1, and, consequently, (5) do not provide correct incentives, which, in a nutshell, means that (6) Draghi's "whatever it takes" is a contradiction to Draghi's Report.
the bit we keep noticing on our side (we track what vcs are funding) - the founders in tallinn, warsaw, vilnius behave very differently from those in paris and milan. do you think a fiscal crisis in italy or france would actually unstick things, or do the distributional coalitions just absorb the shock and keep moving?