European hourly productivity, measured in purchasing power parity at current prices, has remained stable relative to American productivity for twenty-five years. This is the figure Paul Krugman put on the table to argue that the alarm raised by Mario Draghi about European decline was, if not false, at least poorly framed. It’s a good figure. And it’s the wrong indicator.
Philippe Aghion, Antonin Bergeaud, and Luis Garicano published their response on May 29, 2026 in Project Syndicate. It is short, precise, and it has massive political consequences. Their argument: purchasing power parity at current prices incorporates differences in price levels between countries, which creates a systematic bias in favor of economies where non-tradable services—health, housing, public education—are structurally cheaper. Correct this bias, take productivity at constant prices or measure it in sectors exposed to international competition, and the gap between Europe and the United States does not disappear. It widens.
Two first-rank economists. The same twenty-five years of data. Opposite conclusions. This is not an academic quarrel. It is a disagreement about reality that, at this precise moment, is orienting investment decisions, market reforms, and industrial policies representing trillions of euros.
The Essentials
- Krugman relies on hourly productivity in purchasing power parity at current prices: stable relative to the United States over twenty-five years.
- Aghion, Bergeaud, and Garicano (Project Syndicate, May 29, 2026) respond that this indicator incorporates a price bias from non-tradable services that masks a real and growing gap.
- The stakes are not academic: the Krugman reading invalidates competitiveness reforms; the Aghion reading makes them urgent.
- Diane Coyle had documented it: current statistical instruments were designed in the 1940s for an industrial economy. They poorly capture services and innovation.
- If Europe pilots its industrial policy on the wrong indicator, Aghion’s theory of growth predicts structural technological decline in twenty or thirty years.
Purchasing Power Parity at Current Prices, a Distorted Mirror
Krugman’s argument rests on a real observation: in current PPP, hourly productivity in France or Germany has not lagged relative to America’s. This is true in OECD statistics. The conclusion he draws from it—that the alarmism of the Draghi report is exaggerated—is plausible if one accepts the indicator.
Aghion and his co-authors refuse to accept the indicator. The technical problem is as follows: PPP at current prices is calculated by including the prices of all goods and services in the economy, including those that do not trade on international markets. Now non-tradable services—hospital care, public housing, administrative services—are structurally cheaper in Europe than in the United States for institutional reasons that have nothing to do with productivity. A French general practitioner is not less productive than an American doctor because a consultation costs 30 euros instead of 200 dollars. The price difference reflects choices in regulation and public financing, not a difference in productive efficiency.
When these prices are aggregated in the PPP, European productivity in exposed sectors is mechanically inflated. The indicator gives the impression of a convergence that does not exist in the sectors that matter for international competitiveness.
The correction proposed by Aghion, Bergeaud, and Garicano is methodologically standard: measure productivity in sectors directly exposed to international competition, or use constant prices rather than current prices to neutralize price structure effects. Result: the gap with the United States is visible, and it has widened since the turn of the 2000s, a period that coincides precisely with the acceleration of the American digital revolution.
What Coyle Saw Before Everyone Else
Economist Diane Coyle has been working for years on what economic statistics can and cannot measure. Her central thesis: national accounting systems were built in the 1940s to meet the needs of an industrial economy—to measure physical production, hours worked, tangible capital. They poorly read an economy dominated by immaterial services, network effects, data value, and innovation quality.
GDP counts a software application downloaded for zero dollars as a null contribution to production, even if it massively increases productivity for millions of users. It undervalues innovation when it translates into price decreases rather than volume increases. It does not capture value creation in digital ecosystems where the boundary between producer and consumer blurs.
This point is directly relevant to the Krugman-Aghion debate. If the United States has produced, since 2000, a disproportionate share of the world’s large digital platforms, and if a substantial portion of the value created by these platforms escapes standard measurement instruments, then the productivity comparison between Europe and the United States is doubly biased: by non-tradable prices, as Aghion shows, and by under-accounting of American digital innovation. Both biases work in the same direction: they underestimate the real gap.
This is a problem of statistical governance as much as economics. One cannot pilot an industrial policy with instruments calibrated for another era. The Draghi report, despite all the documentation it provides on European weaknesses, is built on a diagnosis constructed with imperfect indicators. Redoing the statistics is not an abstract prerequisite: it is a condition for political decisions that follow to be based on reality rather than on a distorted photograph.
Two Diagnoses, Two Policies, One Possible Choice
The statistical quarrel would interest only economists if it did not have direct implications for policy choices. It does.
If Krugman is right—that is, if Europe has not fallen behind in productivity—then calls for deep structural reforms rest on a false diagnosis. Europe might adjust at the margins, but it has no urgent business to conduct. The European social model, with its more protected labor markets, its fragmented capital markets, and its national champions protected from creative destruction, is not in question.
If Aghion, Bergeaud, and Garicano are right, the situation is fundamentally different. Europe has allowed a structural lag to accumulate in sectors with high innovation intensity, precisely because it has protected its established players from the competition that alone forces technological upgrading. The theory of growth through creative destruction, for which Aghion received the Nobel Prize in Economics in 2025, predicts exactly this result: an economy that keeps alive firms whose productivity is below the technological frontier slows its own growth, because it immobilizes capital and labor that could be redeployed toward more productive uses.
Europe has not created a technological giant in fifty years—it is not a deficit of initial financing or intelligence, it is a deficit of destruction. The mergers that would have allowed the creation of actors of world scale were blocked in the name of protecting domestic competition. Capital markets, fragmented along national lines, did not allow large-scale financing of scale-ups. Insolvency rules kept zombie companies that Schumpeterian logic should have eliminated much earlier.
The Aghion reading thus leads to a specific program: unblock cross-European mergers in strategic sectors, unify capital markets, reform competition rules so they target real market power rather than the nominal size of players, and accept that creative destruction implies painful sectoral conversions in the short term. This program is politically difficult. It is all the more so because the diagnosis on which it rests is contested.
AI Reshuffles the Cards Before Europe Decides
The statistical debate is unfolding at the worst possible time. Generative artificial intelligence is reshuffling competitive positions in almost every sector of the economy, and at a speed with no comparison to previous technological cycles. Companies that master foundation models, computing infrastructure, and training data will take structural positions that will be difficult to contest in ten years.
Now, in this specific sector, the gap between Europe and the United States is not measured in hourly productivity points: it reads directly in market capitalizations, venture capital investments, patent filings on language models, the location of talent working on the most advanced systems. These indicators are not ambiguous and do not depend on the choice between current PPP and constant prices.
The question Aghion poses about competition policy takes on particular significance here. If Europe continues to artificially fragment its digital markets to protect nationally-sized players from insufficient scale, it deprives its companies of the critical mass necessary to compete with American players whose resources are incomparable. The tension between competition policy designed for the domestic market and the need to create global-scale champions is real, and it has no simple solution.
Aghion’s argument about creative destruction applies here with particular acuity. An economy that protects its established players from digital disruption does not create stability: it accumulates a lag that, after a certain number of years, becomes structurally difficult to close. The theory predicts a point beyond which protectionist economies cease to converge toward the technological frontier and begin to move away from it in a self-sustaining manner.
Can Europe Innovate Without Being Bought Out?
The question that the Draghi report did not settle is that one. It is different from the debate over past productivity, but it is its logical consequence.
If the productivity gap in innovation sectors is real, as Aghion and his co-authors argue, then Europe faces a structural choice: either it creates the conditions for endogenous innovation by accepting the political costs of creative destruction, or it attracts foreign capital and technologies at the risk of growing technological dependence.
Both options coexist in the current debate. The European Union has committed substantial investments in semiconductors with the European Chips Act, in AI with calls for proposals under the Horizon Europe program, in data infrastructure with sovereign cloud projects. These initiatives respond to Mariana Mazzucato’s logic on the entrepreneurial state: public investment as a catalyst where the market does not invest spontaneously.
But these investments are insufficient if market structures remain unchanged. Financing public research laboratories without creating the conditions in which results can be commercialized at large scale by European private companies reproduces Europe’s historical problem: excellence in basic research, inability to scale. The Draghi report documented this by identifying the capital markets union as the priority unfinished business of European integration.
The Krugman-Aghion debate lands here. If European decision-makers accept the Krugman reading, they will conclude that the situation is not urgent and will defer difficult reforms. If the Aghion reading prevails, the sense of urgency changes in nature: we are no longer in gradual catch-up, we are in the prevention of a structural decline whose consequences will be measured in terms of real income, quality of public services financed by overall productivity, and the ability of future generations to exercise their choices in an economy that remains competitive.
Statistics are never neutral. Choosing one’s indicator is choosing one’s diagnosis. And choosing one’s diagnosis is already choosing one’s policy. Europe cannot afford to choose the wrong indicator and discover the error in twenty years.
One question remains that neither Aghion nor Krugman settles: at what horizon is creative destruction politically tenable in European societies that have built their social contract on employment security and the predictability of career paths? The reforms called for by Schumpeterian theory presuppose an infrastructure for protecting transitioning workers that Europe has not yet fully built. That may be the real prerequisite—not statistical, but political.
Sources
- Aghion, Bergeaud, Garicano — “Paul Krugman Is Wrong About the US-Europe Productivity Gap”, Project Syndicate, May 29, 2026: https://www.project-syndicate.org/onpoint/paul-krugman-is-wrong-about-us-europe-productivity-gap-by-philippe-aghion-et-al-2026-05
- Diane Coyle — GDP: A Brief but Affectionate History, Princeton University Press (data and arguments on the limits of national accounting)
- Mario Draghi — The Future of European Competitiveness, report submitted to the European Commission, September 2024
- OECD — Productivity database and comparative economic analysis (stats.oecd.org)