Improving labor market outcomes of vocational training programs in India: Evidence from an innovative training approach

• In Event: Funding and outcome based decisions in education

Proposal

Overview: A mismatch between the skills that youth possess and the skills that employers require is an important constraint to youth employment globally, including in India. However, the results of many large, publicly funded vocational training programs in India have fallen short of their aspirations.

To provide evidence about how these results might be improved, Mathematica studied the labor market outcomes of trainees in two training programs implemented through a partnership between the government and Generation India Foundation, a non-profit skilling organization. These programs, which were implemented in public training centers across India, prepared trainees for entry-level positions as customer sales representatives in the brick-and-mortar retail industry and as customer service call center operators. Generation India’s innovative training approach included careful selection of highly motivated and committed trainees, high-quality instruction using in-person or blended (mixed of in-person and online) training models, intensive efforts to identify potential job opportunities, and mentorship for trainees.

This presentation will describe our approach to evaluating these two training programs, share key findings on long-term employment outcomes and results from a simple cost-effectiveness analysis, and reflect on the implications of those findings for the wider community of stakeholders interested in improving labor market outcomes of youth through large public vocational training programs, including tech programs.

Approach: We compare the long-term outcomes of Generation learners to those of learners from similar publicly funded programs that used the standard, non-Generation methodology, By carefully selecting the comparison program cohorts to match the training center type, location, and timing of the Generation program cohorts—and controlling for the few differences in trainee socio-demographic characteristics—we can more confidently interpret the differences in labor market outcomes as being driven by Generation’s methodology.

We independently measure the labor market outcomes of Generation learners about 15 months after they completed their Generation programs, on average. These outcomes were measured through a phone survey of 560 learners from 47 training cohorts in India. We measured the outcomes of comparison program cohorts through a phone survey of 509 learners in 115 training cohorts.

Key findings: Generation trainees substantially outperformed comparison trainees in key labor market outcomes related to employment and earnings. Specifically, Generation trainees were more than twice as likely as comparison trainees to have found a job since program completion, and more than 50 percent more likely to be employed at the survey date. Among the sample of those employed, average wages were similar for Generation and comparison trainees. However, in the full sample, Generation trainees earned an average of about 75 percent more at the survey date because a smaller percentage were unemployed (and hence had zero earnings).

We can conclude with reasonable confidence that these positive results are attributable to Generation’s methodology relative to the business-as-usual methodology, including its efforts to select highly motivated and committed trainees, provide high-quality instruction, and proactively seek out job placements. Although Generation’s methodology is associated with higher costs than business-as-usual programs, a simple cost-effectiveness analysis shows that it is about 30 percent more cost effective. The findings provide important evidence for policymakers and other practitioners seeking to drive further improvements in skilling in India and elsewhere.

Authors

• Evan Borkum, Mathematica

• Edith Felix, Mathematica

• Irina Cheban, Mathematica