Global business concentration has risen since 1900, new study finds
New research traces business concentration since 1900, finding revenues, profits and capital increasingly clustered in a few firms across ten market economies.
A broad historical reconstruction by researchers at the Kiel Institute, the University of Chicago and Northwestern University shows that business concentration in revenues, profits and capital has risen steadily over the past century. The study, presented as NBER Working Paper 34711 (2026), uses newly digitized tax and company-count data from ten market-oriented economies to map firm-level shares from 1900 to 2020. While the largest firms now account for a far greater share of turnover and capital than a century ago, the share of employment they control has remained relatively stable. The findings sharpen debate about how technology, particularly automation and recent advances in artificial intelligence, reshape firm size and the labour market.
Scope and sources of the historical reconstruction
The research assembles long-run series from tax records and corporate censuses that have not previously been combined for such extended periods. Data come from ten countries across Europe, North America, Asia and Australia and include early twentieth-century revenue and tax registers that are comparable to sources used in inequality studies. By estimating the shares of the top one percent of firms in sales, profits and capital, the authors create a century-long picture of corporate concentration. These measures allow comparisons across countries and sectors while avoiding the need for exhaustive firm-level microdata.
Rising shares of revenue, profits and capital
Across the surveyed economies the top one percent of companies increased their share of national revenues and capital markedly since the 1920s. In Germany, for example, the largest one percent of firms accounted for roughly half of total revenues in the 1920s; that share is now closer to three quarters. Similar patterns appear for profits and capital holdings, indicating that financial and productive resources have become more clustered in a relatively small number of corporations. The researchers note that this is a widespread, cross-country trend rather than the result of idiosyncratic national policies.
Employment concentration has not kept pace
Contrary to some expectations, the rise in revenue and capital concentration has not been matched by an equivalent increase in employment concentration. Historically and today, the largest one percent of firms still employ roughly half of the workforce in the aggregate samples studied. This gap between revenue and employment shares is particularly pronounced in capital-intensive industries where growth has been driven by investments in machines, software and data-driven processes. The divergence suggests firms can expand turnover without proportionally expanding payrolls.
Sectoral differences and the role of automation
The century-long trends vary by sector: early industrial economies saw large industrial firms rise in importance during the first half of the twentieth century, while trade and services concentrated later, especially from the 1970s onward. Where activities are easier to automate, such as manufacturing, firms tend to grow via capital investments that substitute for labour, increasing revenue and capital concentration more than employment. In contrast, retail and some service sectors have shown parallel increases in revenue and employment concentration, illustrated by the rise of large supermarket chains that now rank among the country’s biggest employers.
Why common explanations fall short
The authors evaluate alternative explanations such as rising market power, outsourcing and global value chains and find them insufficient to fully explain the long-term pattern. Markups and price-setting power rose notably only from the 1980s in many countries, whereas concentration began to increase much earlier. Outsourcing would be expected to leave identifiable signatures in national accounts, but those patterns do not match the historical data. Likewise, the expansion of international production does not account for concentration increases that predate modern global supply networks. Instead, the evidence points toward persistent, economy-wide forces—chiefly technological change—that have reshaped how production is organized.
Policy and labour-market implications in the age of AI
The study raises fresh questions about the labour-market inclusiveness of growth when revenue and capital become concentrated in fewer hands. If firms can boost sales and productivity without hiring proportionally, the integrative effects of growth may weaken. The recent surge in artificial intelligence heightens this concern: AI can scale output and services while substituting for human labour in many tasks, enabling relatively small teams to generate large revenues. The authors point to examples such as legacy software firms that employ tens of thousands while generating billions in turnover, alongside younger AI companies that may reach comparable revenues with far smaller payrolls.
The long-run evidence does not equate size with market dominance in every market, but it does underscore how technological forces have altered corporate footprints. Policymakers and labour-market institutions will need to weigh competition policy, training and redistribution measures against the reality that capital-led growth can expand firms’ economic weight even as employment remains broadly distributed.
