As the world grows richer and technology advances, people should have to work less. That’s been the conclusion of thinkers ranging from John Maynard Keynes a century ago—when he predicted a 15-hour workweek by now—to AI enthusiasts today.
And it’s entirely wrong, according to our research using a new global dataset we developed. We find no evidence that working hours fall for people of prime working age as economic development expands. What does affect hours at work are social choices regarding schooling for the young, pensions for older people, and regulations and income support for everyone else.
These findings have profound implications for economists’ understanding of how the world works, and how much. The common view among economists has been that working hours are an individual decision. In fact, policymakers need to understand that the organization of work reflects a complex interplay of cultural and social choices as encoded in public policies, including labor law, taxation, education, and retirement benefits.
Working time is a fundamental determinant of economic production, well-being, and inequality. Our new global dataset covers nearly the entire world population and provides a clearer view of how working hours vary by age, gender, income level, sector, and institutional context. This database also includes time series spanning several decades in many countries at all levels of development, allowing for a broad understanding of global trends in working time.
Our main data sources are household surveys fielded by statistical institutes and compiled by the International Labour Organization and the World Bank. These surveys record detailed information on hours worked, together with data on age, gender, sector of employment, and earnings.
We supplemented the data with other surveys from various international and country-specific sources. The resulting publicly available database includes almost every labor force survey ever fielded in the world. It covers 160 countries representing 97 percent of the world’s population. Our time series span more than 20 years in 86 countries at all levels of development.
In compiling our dataset, we follow international conventions, measuring weekly hours worked in all jobs that contribute to economic production. We thus include unpaid agricultural work, which produces goods and counts in standard measures of economic output, but we exclude unpaid home services such as cleaning, cooking, and taking care of children or elderly family members.
We find that globally 59 percent of adults ages 15 and older are employed. Those with jobs work 43 hours a week on average. This implies that the average adult worldwide—including those without employment—works about 25 hours each week.
This global figure hides sharp variations by age and gender (see Chart 1). Working hours follow a pronounced life-cycle pattern. They are lowest among teenagers, rise steeply during early adulthood, peak in prime working years, and fall sharply after age 60. Gender differences continue to be striking. Men provide roughly two-thirds of total worldwide working hours, and women account for just one-third.
These disparities arise primarily from differences in employment rates rather than differences in hours among those already employed. The core divide is not mainly about how long people work, but who participates at all.
It is often thought that people in richer countries work less. The reality is more nuanced. Hours worked per adult follow a mild bell-shaped relationship with income per capita. They are lowest in poor and rich countries and highest in middle-income economies. Yet income levels explain only a small fraction of global variations in working hours (Chart 2).
Some of the lowest working times are in countries as diverse as France, South Africa, and Afghanistan; some of the highest are in Madagascar, Vietnam, and China. Development alone cannot explain why societies with similar income levels work vastly different hours.
The bell-shaped relationship between hours and development is driven by hours per worker rather than by employment rates. Working time for those with jobs exceeds 45 hours a week in middle-income countries such as India and Pakistan.
Structural transformation plays a key role. In emerging market economies, rapid expansion of manufacturing and services coincides with intense labor demand and long working hours. In high-income countries, stronger regulations, greater formalization of work, and shifting social norms gradually reduce weekly work time, even as employment remains high.
There are major exceptions to this pattern, however, such as the long hours worked in rich countries, including the United States and Singapore. Development expands capacity and productivity, but it does not mechanically dictate how much people work.
Working time among young and elderly populations declines sharply as countries develop. This might appear to reflect rising wealth and preferences for leisure. In practice, it has to do primarily with institutional change. For young people, expanding school attendance results in declining work hours. As countries invest in education, adolescents shift from working to studying. For older adults, reduced working time comes with the development of public pension systems, which allow retirement without severe income loss.
Once we take schooling and pension coverage into account, national income itself no longer predicts working hours at these ages. Young and elderly workers in richer countries work less not because of higher incomes but because governments decide to invest more in education and pensions.
Among prime-age adults (those who are 20–59 years old), total working time has been remarkably stable across development levels and over long historical periods. In the United States, for example, average prime-age hours, including those without jobs, were about 30 a week in 1900—virtually identical to today. A notable exception was a sharp drop in working hours during the Great Depression. The data show similar stability in western Europe, Latin America, sub-Saharan Africa, and many parts of Asia.
However, beneath this stability lies a deep reorganization of work by gender. As countries develop, male working hours decline and female hours in paying jobs increase. In many regions of the world, these two forces exactly offset each other.
For men, this change reflects a decline in hours on the job. For women, it reflects rising employment rates as barriers to participation fall. Development therefore does not reduce total prime-age labor. Rather, it redistributes the work across genders, leading to a long-term reduction in gender inequalities in working time.
We do observe a reduction in working hours in many countries at later stages of economic development. This is often related to the development of welfare states discouraging labor supply, providing an explanation for the higher hours worked in the United States compared with Europe.
Across countries, we find that higher labor taxes correlate with fewer working hours. Countries with low taxation, such as India, tend to record longer hours, while people in high-tax countries, such as France, work considerably less. At first glance, this appears to confirm the view that taxation discourages labor supply.
A closer look reveals a more complex reality. Once we account for labor regulations and the expansion of formal employment, the direct effect of taxes on working hours disappears. High-tax systems coincide with stronger regulation of working time and broader coverage of formal labor contracts. These appear to play the most important role in explaining the decline in hours at the highest levels of development.
Reduced hours therefore reflect policy choices about how work should be organized, not simply individual responses to incentives. Limits on maximum working hours, overtime rules, paid leave, and retirement systems all shape national patterns of work much more strongly than taxes alone.
These findings suggest that cultural and social choices encoded in public policy shape hours worked much more than pure economic development. The balance between work and leisure is not simply about personal preference. It is a collective negotiation embedded in institutions, labor laws, and social expectations.
AMORY GETHIN is a researcher in the World Bank Development Research Group and a coordinator of the World Inequality Lab at the Paris School of Economics.
EMMANUEL SAEZ is a professor of economics and director of the Stone Center on Wealth and Income Inequality at the University of California, Berkeley.
