M Abdullah bin Zubair Hayat
Pakistan is a young country that measures its young late and loosely. The Pakistan Bureau of Statistics (PBS) reports that 58.3 percent of the population is below the age of 25, with 46.3 million aged 15 to 24. Even that count is an extrapolation. PBS grows the 2023 census figures at an assumed annual rate while official population projections are still being finalised, and the labour force survey before the latest one was fielded in 2020/21. A country adding millions of workers cannot steer by a four year old map.
What the map shows is sobering. Between the two rounds the labour force grew from 71.8 million to 85.6 million, about 3.5 million people a year, while the number of unemployed rose from 4.5 million to 5.9 million. Unemployment among 15 to 24 year olds climbed from 11.1 to 12.5 percent on the comparable definition. A companion PBS report finds that only 38.7 percent of that age group is employed, and that 28.4 percent are not in employment, education or training, a rate running from 24.8 percent in Punjab to 36.0 percent in Balochistan.
The deeper lesson is that joblessness understates the problem. About four in five jobs are informal, though PBS cautions that its latest round adopted new international standards, so part of the movement reflects measurement. Education offers uneven protection. Unemployment among female degree holders is 24.1 percent against 6.8 percent for men. Upstream, the World Bank’s Human Capital Index says a child born in Pakistan will be 41 percent as productive as she could be, and about three in four ten-year-olds cannot read a simple text. Later World Bank estimates put that share higher still. UNICEF counts over 26 million school-age children out of school. These are the numbers that become tomorrow’s youth statistics.
The strongest evidence that a youth bulge can pay comes from East Asia. Bloom and Williamson attributed one-third to one half of the region’s growth between 1965 and 1990 to its demographic gift, while stressing that the gain depended on policies that put the larger working-age population to use. They projected that Pakistan and Bangladesh would gain most in South Asia. Bloom later described Latin America as an example of an unreaped dividend. On measurement, Mexico’s Progresa school subsidy was randomly allocated across 495 localities, which allowed researchers to estimate its effect cleanly, and they found a consistently positive impact on school continuation. Evaluation was designed in before the money moved.
Two changes would matter most here. The first is to link school, training and employment records, with privacy safeguards, so a district can see where young people fall out of the system. The second is to allocate money by need and test every youth programme against a comparison group. With provincial NEET rates spanning 11 percentage points, a formula based on population alone is blunt.
Artificial intelligence raises both the promise and the stakes. The National AI Policy 2025 aims to train 1 million professionals by 2030 and launch 50,000 AI-enabled civic projects. A target is not an outcome, but the tools are real. Early warning models for dropouts and skills matching tied to local demand are well within reach of a state that has its records in order. The ILO finds that one in four workers globally is in an occupation with some generative AI exposure, only 3.3 percent are in the highest category, and transformation of jobs is the likelier effect. Exposure is 11 percent of employment in low-income countries against 34 percent in high-income ones. That is hardly consolation if it reflects a thin digital economy.
There is also the question of who stays to benefit. The Bureau of Emigration registered 762,499 workers for overseas jobs in 2025, 5.1 percent more than the year before. About 70 percent went to Saudi Arabia. Set against 3.5 million new labour force entrants a year, that is roughly one in five. An AI strategy that trains a million people and then exports them has outsourced the problem.
The future will depend less on algorithms than on habits. Publish on a fixed schedule, link the records, fund by need and evaluate honestly. A young Pakistani born today deserves a government that can count them and then act on the count.
—The writer is a student of Data Science in Public Policy at the London School of Economics (LSE)
