Europe’s Postwar Promise Was Never Just About Welfare

In wartime Britain, officials knew that social insurance depended on jobs, medical care and contributions people could afford. Eight decades later, Europe’s disputes over pensions, housing, productivity and AI expose how much its social promise still depends on conditions beyond the law.

On December 2, 1942, with Britain still at war, officials at the Ministry of Labour put a question to William Beveridge's proposed system of social insurance: could the nation afford it? They compared the costs expected in 1945 with those projected for 1965, weighed the burden on employers and anticipated the demands of an aging population. Other questions cut closer to the lives the plan was supposed to protect. Could workers afford their own contributions? Would jobs still be available when wartime production ended? The document was not an attack on the idea of social security. It showed how seriously its architects had to consider the economic arrangements on which a promise of security would depend.

Beveridge's report, Social Insurance and Allied Services, published that December, addressed five forms of deprivation: want, disease, ignorance, squalor and idleness. Its contribution-based insurance scheme would provide a basic income when earnings stopped. But Beveridge understood that income protection had limits. Family allowances, comprehensive medical care and the maintenance of employment were explicit assumptions of his plan. The Ministry's officials were particularly troubled by employment: industries leaving wartime production could face new competition while employers resisted higher charges on wages. The government could establish an entitlement. Sustaining the conditions in which that entitlement protected a working life required more than a statute.

There was public appetite for this undertaking without universal confidence that it would happen. A wartime opinion survey preserved in the British archives recorded strong support for the report's provisions, including health care available to everyone, while also finding considerable doubt about the government's willingness to implement them. The surviving evidence belongs to that particular survey, not to a timeless national mood. Even so, it corrects a seductive memory of the postwar years. People did not need to believe that the future was secure before asking institutions to make life less precarious.

The problem of making social obligations credible has not disappeared. On October 23, 2026, France's National Assembly rejected the revenue section of a bill to finance social security in 2027. Under French legislative rules, the bill was consequently rejected at first reading without a separate vote on its expenditure section; its ultimate fate was not settled by that vote. In another part of European policymaking, the EU had revised the timetable for applying some of its safeguards to high-risk AI, including systems used in employment. These were separate decisions, arising from different circumstances. Together they bring into view two kinds of obligation: maintaining protection that citizens already rely on and making new protection enforceable as work changes. The adequacy of either depends on institutions and economic conditions beyond the wording of the right itself.

There was no clean passage from the report to a finished welfare state. Britain already had forms of social insurance before the war. After it, political and legislative decisions translated parts of the new ambition into institutions. The National Health Service Act of 1946 provided for a comprehensive health service in England and Wales; the NHS began operating in July 1948. Insurance, health care, assistance and the place of local authorities developed through different decisions. The result was transformative in its reach, but it was made by politics and administration, not by the publication of a report.

Nor did Western Europe adopt a single British design. France’s October 1945 ordinances organized social security around contributions from employers and workers and institutions with a role for their representatives, while retaining distinctions among existing schemes. German social insurance had roots in the nineteenth century. The postwar settlement drew on older systems even as governments extended their reach and altered the relationship between paid work, family life and public obligation. There was no moment when Europe agreed on one definitive formula for a good society.

Five years after the war, another proposal approached the future through the material foundations of conflict. On May 9, 1950, French foreign minister Robert Schuman called for French and West German coal and steel production to be placed under a common authority. The industries essential to rebuilding Europe were also essential to making war. Joint oversight was intended to make renewed conflict between the two countries less feasible while modernizing production and improving economic conditions. Schuman did not claim that Europe could be remade in one stroke. He proposed beginning with a limited institutional arrangement and allowing cooperation to grow from its practical consequences.

The British insurance proposal and the French plan for industrial cooperation addressed different dangers. One set out ways to protect income when work stopped; the other placed strategic production under joint oversight to make a return to Franco-German war less feasible. Neither created an identical European model of society, and neither institution by itself explains the peace or prosperity that followed. Yet each made a concrete institutional commitment where governments might otherwise have relied on hope: states would pay benefits according to rules, and states with a history of fighting one another would manage vital industries together. Such commitments could change behavior only to the extent that the arrangements survived political conflict and their economic foundations proved workable.

That history cannot be told as a universal emancipation. Beveridge’s architecture assumed a pattern of full-time male employment and often treated married women’s claims through their husbands’ contributions. The 1950 Schuman Declaration, for all its language of peace and cooperation, also described the development of Africa as a European task. When the Treaty of Rome was negotiated in 1957, colonial relationships shaped which people could obtain the new Community’s legal benefits and representation. These were not minor imperfections discovered only by later generations. They marked the boundaries of the political community whose future was being organized.

The exclusions belonged to the history of this achievement. Protection against illness, unemployment and old age became more reliable for millions, while access and representation remained shaped by gender, occupation, citizenship and empire. Later generations inherited functioning forms of security along with rules written for particular kinds of work and family life. Expanding their reach would require political decisions and economic resources, not simply reverence for the original settlement.

What followed provides evidence of an extraordinary improvement in everyday life. It also explains why the gradual disappearance of the postwar boom should be analyzed as a change in the conditions of progress, rather than as proof that the progress itself was illusory.

The Conditions of Prosperity

The institutions taking shape after the war were accompanied by changes in daily life too substantial to dismiss as political mythology. In metropolitan France, 52 infants died before their first birthday for every 1,000 live births in 1950. By 1973, the rate had fallen to 15.4. Over the same years, life expectancy at birth rose from 63.4 to 68.7 years for men and from 69.2 to 76.3 for women. These figures do not isolate the effects of social insurance from those of medicine, sanitation, nutrition or rising incomes. They establish something more basic: the decades associated with the expansion of postwar social protection also brought measurable improvements in the chances of surviving childhood and living longer.

The economic transformation was of comparable magnitude. In Angus Maddison's reconstruction, real output per person in 12 Western European economies grew by an average of 3.9 percent a year between 1950 and 1973, against 2.5 percent in the United States. Output per hour worked grew even faster, averaging 4.8 percent annually. At the beginning of the period, Western Europe produced about 44 percent as much per hour as the United States, measured in comparable purchasing-power terms. By 1973 it produced about 68 percent. European workers did not suddenly become the world's most productive. They were narrowing an unusually large gap at extraordinary speed.

Much of this acceleration came from changing where and how people worked. As agricultural employment contracted, workers moved toward industry and services, where machinery, infrastructure, scale and organization could generate more output. Firms invested in equipment and acquired production methods already in use across the Atlantic. Jonathan Temple's research on structural change treats the reallocation of labor as a substantial contributor to the era's aggregate productivity gains. Nicholas Crafts and Gianni Toniolo also emphasize the exceptional scope for catching up from a relatively low postwar starting point. Reconstruction mattered, but rebuilding damaged factories was not the whole story; there were gains to be had from reorganizing an economy that had fallen behind the technological frontier.

An expanding industrial economy required labor, and the people supplying it were not always those imagined in the welfare settlement's original household model. West Germany entered agreements to recruit workers from Italy in 1955 and, later, from several other countries. Its foreign workforce rose from roughly 330,000 in 1960 to 2.6 million in 1973. These workers contributed to production and earned wages that could be attractive relative to those available at home. But the recruitment system initially treated many of them as temporary, even as employment, family relationships and settlement took root. Some lived in tightly regulated dormitories; work arrangements and residence rights did not necessarily offer the security available to established citizens. Women were recruited too, although the public image of the “guest worker” was often male.

The foreign workers recruited into West German factories make the postwar bargain more complicated than an agreement among employers, unions and an already settled body of citizens. Their labor helped meet demand; their position depended on rules governing residence, families and employment that did not always match the security promised to established citizens. Economic expansion could broaden the resources available for protection while leaving unsettled whose contribution entitled them to its benefits.

The pace of expansion had several reinforcing causes. Firms could invest in plants and equipment with established technologies still to adopt; agricultural workers entered higher-productivity employment; education and trade widened the capacity to produce. Barry Eichengreen has argued that wage bargaining and international cooperation also made investment more credible. Employers could plan against a relatively stable industrial environment, while workers accepting wage restraint expected a share of future gains through wages and public provision. This is an explanation of how political arrangements may have sustained growth, not evidence that one institutional design worked equally well everywhere.

The Marshall Plan belonged to this network of conditions. American assistance supplied dollars and strategic support during reconstruction. In a historical reassessment, Bradford De Long and Eichengreen argued that its direct financing was too small to account for Europe's recovery principally through imported machinery or rebuilt factories; its influence on economic cooperation and policy arrangements was more important. Domestic investment, European payments cooperation, expanding trade and the Cold War security order also shaped the environment in which governments and businesses acted. None of them could be isolated as the sole source of the boom.

The fiscal side of this transformation was equally varied. Historical estimates compiled for the OECD show public social expenditure, excluding education, rising from 11.3 percent of GDP in France in 1950 to 16.7 percent in 1970. In the Netherlands, the corresponding share rose from 6.9 to 22.5 percent. Britain's increase, from 10.9 to 13.2 percent, was much smaller. These series encompass different national institutions and are historical reconstructions, not directly comparable to today's harmonized EU accounts. But they undermine the idea that a single welfare blueprint expanded at the same pace throughout Western Europe.

Industrial expansion made more social provision affordable: earnings and output enlarged the contributions and tax receipts available to fund services and benefits. Those services, in turn, changed the risks confronting people who fell ill, lost a job or retired. Better health and a more secure household budget could help people participate in the economy, although the contribution of each program to subsequent growth cannot be extracted from spending and GDP trends alone. Welfare systems also influenced incentives to work and save and the costs faced by employers. What mattered historically was the interaction, not an automatic formula in which growth produced social rights or social rights produced growth.

This prosperity created political room for bargains whose costs had been debated from the start. It also benefited from circumstances that no government could promise to reproduce indefinitely: an enormous technological distance from the American frontier, scope to reallocate labor from lower-productivity occupations and an expanding world trading system. The social gains were real. Their exceptional rate of expansion did not constitute a permanent forecast for growth or tax revenue.

Once much of the easiest catching-up had been achieved, growth would depend more heavily on new forms of investment and organization. The institutions built during the boom would have to operate in economies that no longer offered the same rapid expansion of their financing base.

When Growth and Progress Diverged

The first change came before the digital revolution. By the early 1970s, the gains from rebuilding capacity and introducing already proven production methods had begun to narrow. Oil-price shocks raised costs across industries reliant on imported fuel. Inflation, weak demand and unemployment unsettled wage bargains negotiated during years of fast productivity growth. Governments faced competing claims that rising output had previously made easier to accommodate. But the break after 1973 did not mark a single turning point for every measure of European economic performance.

The statistics expose a history that the conventional date of Europe's decline obscures. Maddison's reconstruction puts average annual growth in output per person across 12 Western European economies at 3.9 percent from 1950 to 1973 and 1.8 percent from 1973 to 1998. But European output per hour continued to gain on the United States through the later period, reaching about 83 percent of the American level by 1998, up from 68 percent in 1973. Per capita output, by contrast, went from 73 percent of the American level to 72 percent. The numbers can be reconciled: output per person equals output per hour multiplied by hours worked per resident. In Maddison's tables, hours worked per resident fell from about 750 to 657 across the 12-country group between 1973 and 1998; in the United States they rose from about 704 to 791. These are reconstructed national aggregates, not measures of the hours worked by a typical individual or of the quality of those hours. Yet they show why the slower convergence of living standards requires an explanation of employment, demography and time worked alongside technological efficiency.

The difference has consequences for how progress is measured. Paid leave and shorter working schedules can give people time for family, care or rest—benefits that GDP per resident records only as production not undertaken. Unemployment and an inability to obtain enough paid work impose something very different. The same numerical decline in hours per person can reflect a welcome increase in leisure for some and lost earnings or exclusion for others. Any assessment of Europe's economic model has to keep those experiences apart.

What turned slower growth into long-lasting joblessness was contested even as it unfolded. Olivier Blanchard and Justin Wolfers examined unemployment in 20 OECD economies from the 1960s onward. They found that exposure to economic shocks did not, by itself, explain why national unemployment trajectories diverged, while the existence of particular labor-market institutions could not, on its own, explain why joblessness rose when it did. Their evidence pointed to interactions between shocks and institutions. The protections and bargaining arrangements that operated in a rapidly expanding industrial economy could produce different adjustment paths when firms faced weaker demand, altered energy costs or slower underlying productivity growth.

Countries responded differently. Wage bargaining, job protection, unemployment insurance and retraining influenced how people and firms absorbed shocks, while monetary policy, investment and demand shaped the opportunities available to them. Sweden and Denmark changed parts of their labor-market or pension arrangements without abandoning social provision. Such adjustments had distributional costs and did not guarantee success. Their significance is that institutional continuity did not mean keeping every rule unchanged. A credible account of European unemployment must examine the effects of particular policies under particular conditions rather than treating total welfare expenditure as a sufficient explanation.

A second turning point emerged in the mid-1990s, long after the first oil shock. The United States began drawing new productivity gains from information and communication technologies, while much of continental Europe advanced more slowly. The difference was not confined to the companies manufacturing computers. It became visible in the businesses using them. Wholesale distribution, retailing, finance and business services could combine software, logistics, data and new ways of organizing work; in parts of the American economy, these changes generated large gains in output from a given hour of labor. European firms also adopted computers. The harder distinction was whether the technology was accompanied by investments and organizational changes that altered how entire sectors operated.

The European Central Bank's retrospective comparison provides a sense of the scale. Between 1995 and 2005, hourly labor productivity in private market services rose by about 1.0 percent annually in the euro area and 2.9 percent in the United States. In wholesale and retail trade, the respective rates were about 1.4 and 4.6 percent. These were historical comparisons for specific sectors and economies, not measures of every European firm's performance. They locate an important part of the divergence in activities that connect producers with customers and finance with investment, rather than in the disappearance of European industrial production.

The relevant innovation was often organizational. A retailer could install computers without altering its purchasing, deliveries or inventory decisions; better information mattered most when it changed those decisions throughout the business. Software could help finance and business services handle more transactions, yet the size of the gain depended on skills, management and complementary investment. Research by Bart van Ark, Mary O'Mahony, Marcel Timmer and their colleagues points toward differences in ICT investment and sectoral efficiency. It does not justify attributing the American advantage to the absence of one European regulation. How technology spreads through ordinary firms matters at least as much as the capabilities of the firms that develop it.

The financial crisis then imposed a third kind of constraint. The collapse in investment during the 2008–09 recession affected both the United States and the euro area, but their recoveries separated when the euro area's sovereign-debt crisis intensified in 2011. Borrowing conditions tightened in the countries under stress, fiscal policy became more restrictive and uncertainty increased. The euro area entered another recession, and investment fell again. This did not originate the technology-related divergence visible since the 1990s. It limited the capital formation and economic experimentation through which a mature economy might respond to that divergence. The distinction matters because a diagnosis of slow innovation and a diagnosis of interrupted investment call for overlapping but not identical explanations.

Nor did the newer European Union follow one downward trajectory. The countries and regions of Central and Eastern Europe that joined the Union from 2004 experienced substantial convergence in income per person. The European Commission's Ninth Cohesion Report put the region at around 52 percent of the EU average in 2004 and nearly 80 percent in 2021, using purchasing-power comparisons. That progress cannot be assigned entirely to a single EU funding program: investment, trade, institutional changes and a lower starting level also mattered. At the same time, parts of Southern and Western Europe struggled to converge or recover from the financial crisis. A history written only through Paris, Berlin and Brussels would miss Europeans for whom integration and industrial development still brought a substantial catch-up in living standards.

The longer divergence remained important even as particular years improved. In the OECD's historical comparison, EU output per hour stood at about three-quarters of the American level in 2024, compared with roughly nine-tenths in 2000, using a defined constant-purchasing-power series. The OECD's preliminary estimates also suggest that the EU's hourly productivity growth accelerated to 1.4 percent in 2025 after 0.2 percent in 2024. The American rate was estimated at 1.7 percent in 2025. These figures relate to the EU-27 rather than Maddison's historical 12-country West European group, so they are evidence of a later problem rather than a continuation of precisely the same national series. The central difficulty is to explain how new productivity gains spread across firms and jobs, and what happens to their proceeds—not to suggest that every European industry is producing less than it once did.

The result is a different account of Europe's subsequent economic difficulties. The postwar boom, the rise of mass unemployment, the diffusion of digital tools and the euro-area debt crisis changed the constraints on firms and households in distinct ways. Even within the Union, the gains from integration were distributed unevenly across countries and regions. To judge whether the institutions inherited from the boom remain effective, the next inquiry has to leave aggregate output behind long enough to examine what people are actually protected against—and which forms of opportunity they can reach.

Security and the Geography of Opportunity

The protective institutions have not stopped working. In 2025 the EU employment rate for people aged 20 to 64 reached 76.1 percent, while unemployment stood near 6 percent. According to the European Commission, real wages had returned to at least their 2019 level in 21 of the 27 member states. These developments arose from many economic forces; they cannot be credited to social insurance alone. But they establish the setting in which the debate over Europe's social model must take place: a society with functioning protection, substantial employment and continuing inequalities, rather than one in which the postwar settlement has simply ceased to deliver.

Consider what benefits do to the distribution of income. The EU's 2025 household survey, based largely on incomes earned in 2024, found 16.3 percent of residents below the relative income-poverty threshold. The corresponding measure before social transfers other than pensions was 24.4 percent. The difference of 8.1 percentage points demonstrates how much the accounting of disposable income changes once those transfers are included. It is not a forecast of what poverty would be if benefits were abolished: people, employers and governments would respond to a different set of rules. Existing transfers nonetheless make a material difference to the income with which millions of households live. Poverty persists despite them, and the ability to escape it depends on conditions beyond the benefit system.

Financing that protection poses a separate question. In 2025 France's general-government expenditure reached 57.3 percent of GDP, against 49.4 percent for the EU as a whole. Its government also collected more: revenue amounted to 52.2 percent of GDP, compared with 46.3 percent across the Union. France still ran a deficit of 5.1 percent of GDP, two percentage points larger than the EU aggregate. These figures show the scale of its collective commitments and the resources already raised to meet them. They do not identify which specific program should change. An expenditure can be valuable and still require a sustainable means of payment; an impressive revenue ratio can coexist with a serious fiscal imbalance.

France's pension projections illustrate the importance of time and accounting. The country's Pensions Advisory Council estimates expenditure of 14.1 percent of GDP in 2025 and roughly 14.2 percent in 2045 in its 2026 reference scenario. Revenue is projected to decline from around 13.9 to 13.3 percent of GDP over that period, in part under financing rules for civil-service and special schemes whose transfers must also be understood within the wider public budget. Beyond 2045, the spending path becomes less favorable: expenditure reaches 15.3 percent of GDP by 2070 under the report's demographic assumptions. These are conditional projections, sensitive to policy and population changes. They reveal why an argument about the cost of aging cannot be reduced to one expenditure number, and why moving a payment between public accounts is not the same as reducing the state's obligation.

A pension is a claim whose financing and value unfold across a working life and retirement. Housing is an immediate condition of living and taking work in a particular place. The distinction matters because a society can preserve a reliable retirement income without making a new home attainable to someone entering a labor market. Conversely, constructing more dwellings does not balance a pension fund. Measuring the quality of a social settlement requires looking at both kinds of capacity: what protection people can count on over time and what opportunities remain accessible when their lives change.

At first sight, European housing figures point in opposite directions. In 2025, 7.7 percent of EU residents were living in households spending more than 40 percent of disposable income on housing, down from 8.2 percent a year earlier. Yet nominal house prices had risen by approximately 62 percent since 2015, generally outpacing income growth after 2016. Existing tenants and owners, people sharing a household and those looking for their first home face different prices and constraints. A household with an established mortgage may have a manageable bill while a would-be resident cannot afford to move into the same city. Improvement in the measured burden of current occupants can coexist with deteriorating access for new entrants.

This is especially important when speaking about younger adults. In 2025, 26.9 percent of EU residents aged 15 to 29 lived in overcrowded housing, against 16.8 percent for the overall population. Yet the Union's average estimated age of leaving the parental home, about 26.3 years, has been broadly stable since 2002. It would distort the evidence to declare a continent-wide, continuously worsening inability to leave home. A person can leave early and struggle to pay rent, or remain longer with family and record a lower measured housing bill. Public support, employment, customs of family life and the local rental market all shape the result. Independent living and affordable housing require more than a single rate to describe them.

A 2026 IMF study of European housing markets found that financially vulnerable renters had become much less likely to move into homeownership than before the global financial crisis. The authors also estimated, through a model of housing costs and mobility, that the EU may have forgone roughly one million moves over a decade. That is a counterfactual calculation, not a count of rejected moves. The mechanism it identifies can be understood without adopting the estimate as an observed fact: a better job in a productive region is worth less to someone who cannot afford to live near it. Legal freedom to move does not pay the rent at the destination.

There is also a value in remaining where one's life is rooted. Care responsibilities, friends and long-established community ties are reasons to stay even when an employer elsewhere offers more money. Renting securely can provide more freedom than an unaffordable mortgage. The policy objective should be to expand credible choices: to make it possible to stay without constant fear of displacement and to move without surrendering every source of security. Mobility is a means available to people, not an obligation imposed in the name of efficiency.

Vienna makes the trade-offs visible because its housing policy supplies a genuine benefit. The OECD reports that 41.7 percent of the city's residents live in social housing. Average monthly rent in those homes was about €550 compared with €807 in the private market, and the scale of provision exerts downward pressure on rents more broadly. That is a substantial contribution to the security of people already living there. At the same time, eligibility tied to a city or region can create costs that become visible when someone considers a job elsewhere. A person leaving Vienna may be unable to obtain comparable support in the new location, even if moving would bring better employment prospects. The OECD identifies this as a potential friction, not an experimentally measured number of missed moves attributable to Vienna's housing system.

This potential cost arises within Austria's uneven geography of work. Unemployment among people aged 15 to 64 stood at about 9.5 percent in Vienna in 2024 and 3.2 percent in Salzburg, where labor markets were tighter. The OECD has proposed expanding affordable housing in regions with jobs and preserving some waiting-list or return rights for people who relocate. These remain proposals, not tested outcomes. Their logic is practical: the security established through one tenancy need not become a penalty when a person accepts work across a regional border. But people who choose to stay in Vienna have not thereby made an economically irrational decision.

Vienna's achievement and its limits have to be understood together. A place-based benefit can protect current residents powerfully while leaving a newcomer—or a former resident trying to return—to meet very different conditions. Extending its reach requires attention to supply and eligibility, rather than simply abolishing a protection because it sometimes affects movement.

Evidence from rental regulation makes that caution necessary. Researchers examining Germany's 2015 rent-brake rules, which limit rent increases on certain new leases, found higher rents in unregulated segments and less residential movement among tenants in regulated units. A study of Catalonia's 2020 rent cap, using rental-contract records and comparisons between municipalities, estimated that average rents fell by 4–6 percent with no detectable reduction in signed or ended leases or the active rental stock in the observed period. Another study of the Catalan intervention found that higher-priced properties left the rental market without equivalent entry at the lower end. Those findings are not interchangeable measures of the same total effect. They examine different segments, observations and identification strategies within a policy that operated for less than two years. The prudent conclusion is that rules aimed at securing existing tenants can have effects on rents and entry that depend on design and supply conditions; the evidence rejects any automatic prediction that every rent cap either succeeds harmlessly or destroys supply.

The studies make a wider point without producing a universal prescription. Housing allowances and some forms of social housing are associated with higher residential mobility in OECD comparisons, while other tenancy arrangements can hold occupants in place. What any rule accomplishes depends on supply, local demand, eligibility and the alternatives available to people facing a change in employment or family circumstances. The object of policy is to improve those alternatives without casually discarding protections that already work.

Such decisions involve people with legitimate but competing claims: tenants who depend on stability, newcomers seeking access, homeowners confronting changes in their neighborhoods and employers needing workers. A city can manage those claims transparently or allow existing rules and prices to settle them without much public scrutiny. The design of housing policy is therefore also a question about whose circumstances count when the meaning of a secure life changes.

Different Economies, Shared Questions

The limits of any one national model become clearer when compared with economies that have made different choices about technology, public provision and development. America, South Korea and China offer no single alternative to Europe's welfare states. Their experiences help locate a common difficulty in separate mechanisms: a country can create substantial wealth without ensuring that everyone can live where its opportunities are generated, receive the resulting income or manage the risks of making use of it.

In the United States, the productivity advantage over Europe is real. Digital investment has helped firms transform parts of the service economy. But higher output per hour has not made access to successful cities easy. The IMF reported in 2026 that average American house prices had risen by more than 50 percent since the pandemic. Expensive borrowing, land constraints and building costs limited the supply response. A skilled worker considering an offer in a fast-growing city confronts the price of housing and commuting alongside the wage. Technology has increased what many firms can produce; land-use and housing institutions help determine who can afford to join them.

For an established homeowner, rising prices may increase wealth. For a prospective buyer without the same assets or family support, they increase the cost of entry. Both experiences can arise from the prosperity of a city; neither alone describes the welfare of the country. This distributional difference is one reason a productivity ranking cannot substitute for an account of where people can live and work.

Nor does productivity dissolve fiscal choices. The IMF estimated a general-government deficit of 6.8 percent of GDP for the United States in 2025. The number comes from a different statistical framework from the Eurostat figures for France, and one year's deficit cannot decide the long-term sustainability of either country. It does show that the world's most productive technology sector can coexist with public borrowing and unresolved arguments about how collective obligations should be financed.

The European Commission's 2026 comparison of selected well-being measures underscores the difficulty of ranking societies by production alone. It found advantages for the United States in household income and employment, while the EU performed better on its reported housing-cost overburden measure and certain measures of personal safety. The comparisons have different definitions, and the costs of private versus public services are not identical. They show why the person weighing earnings, housing, health, security and time cannot be represented faithfully by GDP per hour.

South Korea offers a related but distinct problem: where an economy produces value need not be where that value is received. Its advanced manufacturing clusters in Ulsan and South Chungcheong coexist with the concentration of finance, corporate headquarters and professional services in the Seoul region. The OECD's 2026 survey describes net property-income outflows exceeding 20 trillion won from each of Ulsan and South Chungcheong and inflows greater than 60 trillion won into each of Seoul and Gyeonggi. These are regional income-accounting flows involving dividends, interest and other property income; they are not evidence that a given firm removed the same amount of value from workers or that the flows caused Seoul's housing costs. They expose the spatial separation of production and ownership in a modern industrial economy.

The distinction matters to an individual making a decision about where to live. Seoul offers dense networks of employers, universities and services that can make moving toward the capital economically rational. Yet under the Korean Housing Finance Corporation's stated financing assumptions, only about 7 percent of Seoul homes were affordable to a median-income family in 2025, compared with 32 percent in 2012. This measures the possibility of buying a home with available capital and a conventional mortgage, not the affordability of every rental arrangement or the experience of every prospective resident. Still, it illustrates the difficulty of treating access to the nation's leading labor market as though the market's existence alone made its opportunities accessible.

Moving toward Seoul may raise an individual's pay while leaving the distribution of opportunity largely unchanged. A joint OECD–Bank of Korea analysis of a household panel found that migration to the capital region was associated with higher absolute earnings in many cases, but not necessarily with greater mobility relative to parental earnings. Those from lower-income families were less likely to move in the first place. Selection into migration limits what the study can prove about causation, yet the distinction is vital: higher wages after a move are not the same as a more equal chance of advancement for people who began with different resources.

Korea's public rental housing points to the opposite side of the same problem. Vacancy rates are higher in some developments outside the capital region, where work, transport and services are less accessible. Supplying a dwelling at a lower price does little if it places a household far from the employment and institutions it needs. Vienna shows how location-specific protection can complicate a move; Korea shows how housing built in the wrong relationship to economic activity can fail to attract residents. Housing policy and regional economic development cannot be designed as unrelated projects.

China introduces another mechanism. Massive industrial capacity and investment have not translated into equally strong household consumption. IMF research identifies uncertainty about health care and retirement, limitations in rural social provision, barriers linked to the hukou household-registration system and the need for housing down payments as factors associated with high household saving. When people must retain substantial resources against possible illness or old age, an increase in productive capacity does not necessarily lead them to spend more. Their private decisions respond to the social risks they expect to face.

The IMF has accordingly argued for stronger social protection and more inclusive access to services, alongside reforms to the property sector and local-government finances. Estimates of how much additional rural welfare spending might increase consumption are model scenarios, not results from a completed national experiment. China's political institutions and household-registration restrictions differ fundamentally from those of European democracies. Yet its experience clarifies an economic function of insurance that is easily lost in European budget debates: collective protection can reduce the need for households to guard against every risk through private saving, potentially changing demand as well as well-being.

Together, these cases shift the focus from the amount produced to the conditions for using its benefits. The American case brings land and entry costs into a high-productivity economy. Korea reveals the distance between the geography of production, ownership and family opportunity. China demonstrates how protection—or its absence—can shape the decisions households make with their incomes. Their political and economic systems are too different for the outcomes to be combined into a single formula. They are useful precisely because the mechanisms are distinct.

Europe's coming technological changes will work through these same channels. Productivity growth can produce higher wages, cheaper services or greater returns to owners; employment and investment may concentrate in a limited number of firms and cities. Which outcomes prevail will depend on decisions made well beyond laboratories and software companies. That brings the argument back to the organization of work and to the public institutions that citizens can influence.

Who Gets to Shape the Future?

Artificial intelligence has given Europe's long-standing productivity debate a new urgency. Software can already help with programming, research, writing and routine administrative tasks. But an application that performs a task efficiently is not the same as a firm that reorganizes work around it, and a productive firm is not the same as an economy in which the gains are broadly shared. Europe's experience with computers in the 1990s showed how strongly organizational change and complementary investment can influence the spread of a technology. AI is beginning to pose the same practical questions under different technical conditions.

The differences within Europe are already large. Eurostat found that 20 percent of EU enterprises with at least ten employees used some form of AI in 2025, up from 13.5 percent in 2024. Among large firms the share was 55 percent; among small firms it was 17 percent. Denmark reached 42 percent, while Romania stood at 5.2 percent. These are measures of reported use, not of the sophistication of a company's systems, the value created or the share of its workforce affected. Still, they suggest that any eventual gains may spread unevenly across firms and places that began the transition with very different resources.

Firm-level evidence suggests genuine gains, subject to the limits of the research design. A 2026 European Investment Bank paper matched investment-survey responses with company accounts and used the AI adoption of comparable American firms as an instrument to estimate an effect for European businesses. It reported a roughly 4 percent increase in the level of labor productivity associated with adoption, concentrated in medium and large firms, with no observed reduction in firm employment over the period studied. The instrument's assumptions and the representativeness of adopting firms remain relevant to how far the result can be generalized. It would be inaccurate to translate the estimate into a four-percentage-point annual boost for EU productivity or a guarantee against future displacement. The analysis emphasizes investments in data, software and training that allow AI tools to change company operations.

The International Labour Organization's 2026 review highlights the distance between improvements in particular tasks and gains visible in national productivity data. A document may be drafted faster without shortening the decisions that follow it; an automated stage in a production process may leave the other stages untouched. Firms have to redesign workflows, invest in skills and manage the quality of the resulting work before isolated time savings translate into broader output. Official statistics have not yet revealed a clear economy-wide productivity surge attributable to AI. This leaves both substantial possibilities and substantial uncertainty.

Workers may experience the technology differently even within the same economy. The European Commission found AI use more common in jobs with relatively favorable conditions, while algorithmic monitoring and evaluation extended across a wider set of occupations. For some, new systems may remove repetitive tasks or improve decisions. For others, they can intensify work and reduce control over its pace. Evidence on total employment effects remains mixed. The OECD also notes that deteriorating prospects for some young jobseekers predate the widespread use of generative AI; the technology cannot yet be assigned responsibility for that entire trend.

The way a technology is introduced therefore belongs inside the economic story. In an OECD survey of workers in manufacturing and finance across seven countries, employees at firms that consulted them or their representatives about new technologies were more likely to report benefits from AI. They were nine percentage points more likely to say AI had improved workplace health and safety. That is an association, not a demonstration that consultation caused safer workplaces: employers that consult workers may differ in other important ways, and consultation may alter perceptions as well as conditions. Yet it identifies a practical dimension often missing from forecasts of automated employment. Workers' knowledge, bargaining power and ability to challenge decisions can affect the organization of the technology around them.

Europe is attempting to give part of this process legal form. Article 26 of the EU AI Act requires employers deploying certain high-risk AI systems to inform affected workers and their representatives before use. Information is a safeguard, not a general veto. A July 2026 amendment moved the application of the main obligations for stand-alone high-risk systems, including specified employment uses, to December 2, 2027; the corresponding deadline for certain systems embedded in regulated products is August 2, 2028. The timetable reflects difficulties in preparing standards and enforcement arrangements. Its consequences for innovation and worker protection are still to be observed. For now, the example distinguishes a right enacted in law from a system capable of making it effective in a workplace.

AI also brings the old question of financing back into view. Social protection in Europe draws on employment-related contributions and, in varying proportions, general taxes. A 2026 OECD review identified aging, changing work and the uncertain effects of AI on earnings as challenges for future financing. The resulting fiscal picture depends on whether productivity raises broad labor incomes and employment or shifts more returns toward a narrower group of owners and firms. Technology has not already destroyed the contribution base. But the design of taxes, benefits and protections between jobs will determine how different outcomes affect the capacity to insure social risks.

There is a further difficulty of timing. A company may buy software in months. A worker takes years to acquire qualifications, a household may make a decades-long mortgage commitment, and a pension accumulates over a lifetime. Neither labor-market rules nor housing supply can simply be revised at the speed of a software update. Yet institutions should not become so rigid that the person changing jobs or locations must absorb every cost of adjustment. The challenge is to distinguish predictability, which makes long-term choices possible, from restrictions that protect one situation by closing off others.

Democratic politics has to make these distinctions under conditions of disagreement. Existing tenants and prospective residents may want different housing rules. Employers and workers may have conflicting views about automated performance monitoring. Taxpayers, employees and pensioners have legitimate interests in decisions about future financing. Productivity models can inform those choices but cannot determine whose claims prevail. Elections give governments authority; public explanation, implementation, scrutiny and the ability to revise decisions determine much of what that authority means in practice.

For many citizens, influence over those processes is uncertain. The OECD's 2025 trust survey, published in 2026, found that 68 percent of respondents across participating countries thought voting could have at least a moderate influence on government decisions. Only 31 percent felt the political system gave people like them a say. These are answers to different questions, drawn from participating OECD countries rather than the EU alone. They cannot demonstrate that housing costs or AI caused political distrust. They do reveal a distinction between confidence that elections matter and the sense of being able to affect decisions that shape daily life.

A renewal of Europe's social commitments would have to be judged in the experience of people affected by them. A new technology should be assessed not only by the output of leading firms but by whether other businesses can use it productively and workers can contest harmful decisions. A pension reform should make the distribution of costs and entitlements intelligible. Housing policy should show whether newcomers can enter a city while existing residents remain secure. The point is not to produce a universal list of reforms. It is to make the conditions of economic participation and public protection visible enough to evaluate and change.

The officials who questioned Beveridge's plan in 1942 did not know the shape of the economy their successors would inherit. They nevertheless saw that a legal promise depended on jobs, affordable contributions and medical and family services beyond the insurance scheme itself. The institutions established after the war improved lives, but their rules never reached everyone equally, and the conditions sustaining them kept changing. Those achievements make it possible to distinguish a right from the circumstances needed to use it. A pension requires credible financing; a move toward better work requires somewhere affordable to live; protection against arbitrary decisions at work requires a means of enforcement. Europe's history is evidence of what institutions can accomplish, not a set of instructions that will work unchanged. The future that matters to its citizens will emerge from the economic choices they can make and the public decisions they can meaningfully influence.