Governments and donors have spent billions on vocational training programs across the developing world. The logic seems unassailable: poor people lack skills, employers need skilled workers, so training should connect the two and lift people out of poverty.

Yet the evidence tells a different story. Study after study finds that these programs produce modest effects at best—small gains in employment that often fade over time, earnings increases that rarely justify the costs. Some of the most rigorously evaluated programs show essentially no impact at all.

This pattern raises uncomfortable questions. Why do programs with such intuitive appeal consistently underperform? The answer reveals a fundamental misunderstanding about what actually constrains economic opportunity in developing countries—and points toward more effective alternatives.

The case for vocational training rests on human capital theory. Workers invest in skills, skills increase productivity, and productivity translates into higher wages. When poor people can't afford this investment themselves, public programs step in to fill the gap.

This narrative dominates development thinking. The World Bank has channeled over $9 billion into skills development programs since 2000. National governments spend even more. Training is politically attractive—it suggests poverty stems from individual deficits that can be fixed, rather than structural problems that require harder solutions.

The skills gap framing also appeals to employers who complain loudly about worker quality. Surveys consistently find firms citing inadequate skills as a major constraint. This employer frustration gets translated directly into policy: if businesses need skills, train people in those skills.

But correlation isn't causation, and complaints aren't diagnosis. Employers may cite skills gaps when the real issue is that they can't afford workers with those skills, or when they're unwilling to invest in on-the-job training themselves. The question isn't whether skills matter—they do—but whether training programs are the binding constraint on employment and earnings.

Takeaway

When everyone agrees on an obvious solution that keeps failing, the diagnosis is probably wrong—not the execution.

The most reliable evidence comes from randomized controlled trials that compare training recipients to similar people who didn't receive training. These studies eliminate the selection bias that plagues observational research—the fact that people who seek out training are different from those who don't.

The results are sobering. A systematic review of youth employment programs found that vocational training produces employment gains of about 2 percentage points—statistically significant but economically small. Effects on earnings are even weaker and less consistent.

Some flagship programs perform particularly poorly. Colombia's Jóvenes en Acción, one of Latin America's largest youth training initiatives, showed modest short-term effects that disappeared within two years for men. Kenya's vocational training voucher experiment found no employment impacts after five years. The Dominican Republic's Juventud y Empleo program showed no effects on formal employment or earnings.

The programs that do show positive results tend to share specific features: they combine training with internships or job placement, they're targeted at specific populations like women or the very young, and they operate in contexts with growing labor demand. Even then, cost-benefit analyses often find the programs aren't worth the investment compared to alternatives like cash transfers.

Takeaway

Small positive effects can mask negative returns—if the cost per participant exceeds the lifetime earnings gain, the program destroys value even when it 'works.'

The fundamental issue is that most vocational training treats unemployment as a supply-side problem when it's actually a demand-side problem. Training more workers doesn't create more jobs. If the local economy has positions for 1,000 welders and produces 5,000 trained welders annually, most graduates won't find work in their field.

This insight comes from basic labor economics but gets ignored in program design. Firms don't hire workers because they have skills—they hire because they have profitable work that needs doing. In economies with weak private sectors, limited investment, and few growing industries, skills sit idle.

The displacement effect compounds this problem. When training helps some workers get jobs, they often displace other workers who would have gotten those jobs anyway. From an individual participant's perspective, training helps. From a social perspective, it reshuffles who works rather than expanding total employment.

More promising approaches attack demand directly. Wage subsidies reduce hiring costs for employers. Business training and capital for entrepreneurs creates self-employment. Industrial policy and investment promotion expand the economic pie. These interventions have their own limitations, but they at least address the actual constraint. Training programs keep supplying water when the problem is a blocked pipe.

Takeaway

Supply-side solutions to demand-side problems don't just underperform—they systematically misallocate resources while creating the illusion of action.

Vocational training isn't useless—it's just far less powerful than its popularity suggests. The evidence supports modest, well-targeted programs combined with demand-side interventions, not the massive standalone training investments that dominate development portfolios.

This matters because resources are finite. Every dollar spent on ineffective training is a dollar not spent on interventions that might actually work. The political appeal of training—visible, individual-focused, uncontroversial—makes it hard to redirect funding toward less glamorous alternatives.

Development practice improves when we follow evidence rather than intuition. For employment programs, that means accepting an uncomfortable truth: you can't train your way out of a jobs crisis.