In a nutshell
Vocational training programmes can struggle to have any meaningful and sustained impact on the employment outcomes of the unemployed.
Despite a small short-term improvement in employment in the first year, Turkey’s training programme didn’t improve people’s chances of being employed three years later.
Other evidence indicates that this result is not unique to Turkey, and suggests that in most cases, skills training programmes do not pass a cost-benefit test.
Over the last decade, the World Bank and its client governments have invested almost $1 billion a year in vocational training programmes for the unemployed. But in a large-scale randomised experiment evaluating Turkey’s programme, our research finds that despite a small short-term improvement in employment in the first year, the training didn’t improve people’s chances of being employed three years later.
We conclude that the costs of training the unemployed exceeded the benefits. These impacts are much less than expected by both the participants and the Turkish government.
Unemployment in Turkey rose during the global financial crisis, especially for youth. In response, the government expanded its free vocational training programmes, based on the idea that the unemployed often lack the right skills for the jobs that are available. The courses cover a wide range of fields, they are offered by both public and private providers, and they generally run for six hours a day over a three-month period.
So does vocational training actually improve unemployed people’s chances of finding work, and do the jobs pay enough to justify the costs? The typical problem in evaluating the impact of such programmes is that the people who attend training may not be comparable to those who don’t, making it difficult to know whether any differences in their subsequent employment outcomes reflect the impact of the course, or just that those who take the courses differ from those who don’t.
But in the Turkish case, the courses were so popular that in many cases more people registered than there were available slots. This allowed us to carry out a randomised controlled trial to measure the programme’s impact: 5,902 applicants who passed the eligibility criteria were randomly assigned to a slot in a course (a ‘treatment group’ of 3,001 applicants) or not (a ‘control group’ of 2,901 applicants).
This random assignment ensured that on average the two groups were similar to one another before training, so that any differences in employment outcomes afterwards would reflect the result of the training.
The evaluation covered courses in the last quarter of 2010 and the start of 2011. A follow-up survey took place one year after the courses ended, while employment outcomes over a longer three-year period were obtained by tracking participants through the social security system.
The programme had a small positive impact in the first year. There was a three percentage point increase in formal employment, compared with a 29% employment rate for the control group. But this had disappeared by the third year, when the employment rates were similar for both groups.
The lack of impact doesn’t appear to be due to the duration of the course or the quality of the staff, with no larger impacts for those taking longer courses or courses taught by more educated or experienced instructors.
The cost of the training programme averaged $1,600-1,800 per person trained. People enrolled in public courses earned an extra $15 a month, but this gain did not last more than one and half years, so that the cost of the course wasn’t offset by higher earnings or lower unemployment. Private courses did slightly better, but still had costs that exceeded the gains.
Surveys of programme officials and participants reveal that they expected much higher impacts from the programme: participants expected a 32 percentage point increase in the likelihood of employment, and programme officials a 24 percentage point increase, compared with the short-term increase of three percentage points that actually occurred and no long-term effect.
Our findings suggest that vocational training programmes can struggle to have any meaningful and sustained impact on the employment outcomes of the unemployed. A recent review (Blattman and Ralston, 2015) shows that this result is not unique to Turkey, and suggests that in most cases, skills training programmes do not pass a cost-benefit test.
Blattman, Christopher and Laura Ralston (2015) ‘Generating Employment in Poor and Fragile States: Evidence from Labor Market and Entrepreneurship Programs’, Poverty Action Lab.
Hirshleifer, Sarojini, David McKenzie, Rita Almeida and Cristobal Ridao-Cano (2016) ‘The Impact of Vocational Training for the Unemployed: Experimental Evidence from Turkey’, Economic Journal 126(597): 2115-46.