Original title: An evidence review of worker retraining
David Roodman, Maxim Massenkoff · September 2026
Governments spend large sums of money subsidizing job training to help people build skills and find work. Researchers reviewed major trials from industrial nations and combined data from 56 US experiments dating back to 1973. On average, these programs produce positive but small gains: employment rises 1.7 percentage points in years 3-5, and pre-tax pay increases by roughly $800 annually per person offered training. Big federal initiatives showed tiny or no real effects, and programs only break even under debatable assumptions. However, special "sector programsJob training programs that focus on a specific industry, work closely with employers to design the curriculum, and target local job openings." work much better. They carefully screen applicants, teach skills tailored to local employer needs, and increase earnings roughly ten times as much. Even so, these sector models have proven hard to replicate elsewhere.
You are trying to fix a car engine. You buy a general repair manual, but it is too broad and does not help with your specific model. Then you get a manual written for your exact car, and you fix it easily. In this paper, general job training programs are like the broad manual—they help a little but cost a lot—while sector programs are like the specific manual—they are much more effective because they are tailored to employer needs and local job demands.
Governments spend billions on job training to help unemployed workers. This review shows that most programs barely break even, meaning taxpayer money could be better used. However, sector programs—tailored to specific employers and local demand—are far more effective, raising earnings ten times as much. The challenge is making these programs work at scale.
The meta-analysis covers only US programs since 1973, so findings may not apply to other countries or time periods. The break-even calculation relies on debatable assumptions about how long benefits last. Sector programs have proven difficult to replicate, so their success may not be generalizable. Large federal programs like JTPA and Job Corps show small or null effects, limiting confidence in broad training mandates.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
An evidence review of worker retraining — David Roodman, Maxim Massenkoff
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.