In August 2025, US manufacturing employment was 78,000 below its level from a year earlier; JOLTS openings in the sector stood at 409,000. The IFR analyzes this coexistence in a position paper published in August 2026.
The IFR synthesizes several academic studies published in peer-reviewed journals to settle a forty-year debate over robots’ effect on employment. The answer the report formulates is more precise and more demanding than what that debate had produced until now.
The essentials
- Roboticization often automates tasks rather than entire occupations, but its employment effects also include displacement and depend on productivity and demand effects.
- South Korea has 1,220 robots per 10,000 industrial workers.
- The real bottleneck lies in financing retraining: a displaced worker needs an extended period to retrain, without collective mechanisms to cover that period.
- The IFR forecasts 655,000 global industrial robot installations in 2026, making the question urgent: training and social protection institutions are not calibrated for this pace.
- The political challenge for the next decade is deciding who pays for the transition between the old task and the new one.
The IFR: interested but rigorous arbiter
The International Federation of Robotics is the global lobby for the sector. Its members sell robots. This context must be stated clearly and kept in mind when reading the report. But this positioning does not render the work inept. The IFR chose to rely on several studies published in peer-reviewed journals, not on its own commercial projections.
This methodological choice is verifiable and the studies cited exist independently of the federation.
The position paper published on August 11, 2026 is titled The Impact of Robots: Employment, Productivity and Competitiveness. It responds directly to growing political pressure in several countries seeking to tax or regulate robotic deployments. The IFR wants to debunk what it calls the “myth of job destruction.” The argument deserves to be taken seriously, while noting that certain counterarguments about the distribution of productivity gains are treated superficially.
Robots eliminate tasks, workers keep their jobs
The report’s central thesis is precise: roboticization mainly automates tasks within existing jobs, while its employment effects also include reallocations and shifts between occupations. South Korea’s case is the most telling. With 1,220 robots per 10,000 industrial workers, the country has the highest robot density in the world. Its unemployment rate remains stable. The IFR argues that roboticization can support employment through productivity and demand, without establishing a simple general correlation between robot density and job retention.
The American example complicates the reading, which the report partially acknowledges. In August 2025, the United States had 409,000 job openings in the manufacturing sector. These two figures describe the same phenomenon: companies that roboticize create tasks their current workers do not yet know how to perform. Roboticization transforms tasks and can create a skills gap, making training and retraining necessary.
The IFR also cites productivity data: companies adopting robots can gain in productivity, increase their output and strengthen their competitiveness, which can generate new labor needs in other tasks. The reasoning is consistent with economic literature on creative destruction. It assumes, however, that labor demand remains supported by market growth—an assumption that Carl Benedikt Frey examines with more skepticism in his work on the conditions of progress.
The data the report does not treat sufficiently
The report makes the right diagnosis on employment volumes. It is less convincing on the distribution of gains. The IFR forecasts 655,000 global industrial robot installations in 2026; these deployments generate productivity gains that go somewhere—into margins, prices, dividends, wages. The IFR claims these gains create employment through demand, but without precisely decomposing where they land in company accounts or on what timelines.
This gap does not invalidate the main thesis on volumes, but it weakens the report’s optimism. The academic studies cited show that overall employment does not decline in countries with high robot density. They do not demonstrate that displaced workers quickly find the new tasks created. The gap between these two findings is exactly what unfilled positions in the United States make visible.
China illustrates an amplified version of the same gap. One projection indicates a shortage of 30 million skilled workers across ten critical manufacturing sectors, suggesting that the pace of robotic deployment exceeds the speed of training system adaptation. The pace of robotic deployment seems to exceed the speed of training system adaptation.
Financing the transition
This is the question the report raises without resolving it. A worker whose task is automated does not immediately become competent at the new tasks the robot generates. The sources consulted indicate that professional retraining cycles require an extended period to acquire skills sufficiently different from the eliminated tasks.
During this retraining period, the displaced worker needs income. Existing social protection systems were not designed for retraining of this duration. Conventional unemployment benefits rarely cover more than two years in most OECD countries. Professional training schemes finance short courses, often disconnected from the actual tasks created by companies that roboticize.
Two trajectories open depending on how political decisions are made by 2035. If collective mechanisms finance retraining for the necessary duration, with training and income maintenance, unfilled manufacturing positions could be progressively filled by retrained displaced workers. Several Nordic countries have precursors to such mechanisms, though not specifically calibrated for roboticization. Several Nordic countries have precursors, though not specifically calibrated for roboticization.
Without evolution in financing, the imbalance between unfilled positions and workers without access to retraining could persist. The report presents stable jobs alongside destruction and creation. The World Economic Forum projects a net gain of 78 million jobs globally by 2030, resulting from the ensemble of macro-trends studied. By contrast, territorial inequalities would widen between zones where new tasks concentrate and those where layoffs occur without access to retraining.
The signals that will allow distinguishing between the two trajectories are measurable: the rate of return to employment within twelve months following displacement due to roboticization, compared between countries with a transition fund and those without; and the ratio between unfilled positions and workers in active retraining, measured by sector. This data exists partially in national employment statistics. It is not yet the subject of systematic comparisons at the international scale.
The report’s contributions to the debate
The IFR report has real documentary value: it aggregates several peer-reviewed studies on a subject where literature is scattered and political positions are often hardened before data are read. It clarifies that the overall employment/unemployment debate is the wrong debate. According to the IFR, the effects of roboticization on employment are complex and skills and labor shortages are central concerns.
The right debate concerns the pace of transition and its financing. On this point, the report opens the question without treating it, which is honest but insufficient for a position document meant to guide public policy. There is a missing concrete proposal on retraining mechanisms, and its silence on the distribution of productivity gains leaves the central political question entirely open.
For a reader wanting to understand the state of evidence on roboticization’s impact, this report is a solid starting point. For a public decision-maker seeking what to do about unfilled positions in the United States or the imbalance in South Korea between robot density and training capacity, operational answers will need to be sought elsewhere.
Bibliographic information
Title: The Impact of Robots: Employment, Productivity and Competitiveness Author: International Federation of Robotics (IFR) Publisher: IFR Publication date: August 11, 2026 Document type: Position paper Access: Available on ifr.org
Sources
- International Federation of Robotics, The Impact of Robots: Employment, Productivity and Competitiveness, August 2026
- Bureau of Labor Statistics, Employment Outlook 2024–2034 (BLS, Washington)
- OECD, Automation and Employment Review, 2024
