Improving Youth Employment Opportunities and Reducing Information Barriers through LinkedIn Training in South Africa

Researchers:
Laurel Wheeler
Eric Johnson
Patrick Shaw
Marissa Gargano
Fieldwork by:
Location:
South Africa
Sample:
30 cohorts (1,638 individuals) of job seekers
Timeline:
2016 - 2019
Target group:
  • Job seekers
  • Youth
Outcome of interest:
  • Employment
  • Market access
  • Take-up of program/social service/healthy behavior
Intervention type:
  • Business skills training
  • Information
  • Job counseling
  • Training
  • Certification
  • Computer-assisted learning
  • Employment
  • Hard skills
  • Vocational training (TVET)
AEA RCT registration number:
AEARCTR-0001624
Partners:

Young job seekers in many countries face higher rates of unemployment, underemployment, and unstable employment than older groups, caused in part by information barriers. Researchers conducted an evaluation to test the impact of LinkedIn training on labor market outcomes for young, low-income job seekers in South Africa. Providing LinkedIn training increased end-of-program employment rates by 10 percent (7 percentage points), with effects persisting for twelve months.

Policy issue

Young job seekers in many countries face higher rates of unemployment, underemployment, and unstable employment than older groups. As a result, young workers face high rates of poverty and are more exposed to employment that is non-standard, informal, and less secure than adults. These prevailing issues are caused in part by information barriers in the labor market, such as limited access to referral networks, that impede the ability of young job seekers to enter employment. Although there are several economic factors that may cause high youth unemployment, it may be faster, easier, and cheaper to address barriers to information over other chronic and aggregate economic problems.

Online platforms for networking, job searching, and hiring, such as LinkedIn, have the potential to reduce information barriers and improve employment outcomes for young job seekers. These platforms have also become an increasingly important feature of many labor markets throughout the world, especially for youth seeking high quality employment. 

Does training job seekers in online digital networking platforms, paired with more traditional job readiness training programs, lead to increased employment and platform use?

Context of the evaluation

South Africa has one of the highest unemployment rates in the world, which reached 33.6 percent in 2021. At the time of this intervention in 2017, youth unemployment in the four large cities where this study took place ranged from 39 to 43 percent for those aged 18 to 29. Unemployment is particularly high for young Black people who often face discrimination in the South African labor market.  

During the study period, the Harambee Youth Employment Accelerator (Harambee) ran job readiness programs in South Africa. The job readiness programs were sector-specific and lasted six to eight weeks, focusing on workplace simulations, team building, and non-cognitive skill development. Harambee helped participants submit job applications and set up interviews with sector-specific firms during the program. Study participants were active job seekers, most did not have a university education, few had prior work experience, and most came from low-income backgrounds.

LinkedIn is widely used in South Africa. In 2018, there were 7.1 million active profiles, equal to approximately 40 percent of the workforce, and 264,000 job postings. 

Vocational Training South Africa LinkedIn
Young job seekers completing vocational training in South Africa
Sunshine Seeds, Shutterstock

Details of the intervention

In partnership with Harambee and LinkedIn, researchers evaluated the impact of training young, low-income job seekers to join and use LinkedIn on employment outcomes in four large South African cities. 

Out of thirty cohorts from Harambee’s existing job readiness training program, totaling 1,638 active job seekers, researchers randomly assigned fifteen cohorts to an intervention group that received both Harambee’s standard job readiness program and in-person LinkedIn training, and fifteen cohorts to a comparison group that received only Harambee’s standard program. Of the thirty cohorts, 27 cohorts were geared towards call center jobs and three cohorts were geared towards face-to-face sales jobs in the insurance sector.

The in-person LinkedIn training consisted of an introductory presentation, in-person coaching, discussion sessions, and emails with advice and encouragement that taught participants how to open an account, create a profile, make connections, and ask for recommendations. The LinkedIn training took approximately four hours over the course of the six-to-eight-week job readiness program. 

To measure the impact of the LinkedIn training, researchers conducted surveys before the start of the job readiness program and LinkedIn training, at the end of the job readiness program, and six and twelve months after the program ended. Researchers also used data on platform usage from LinkedIn and administrative data from Harambee. 

This study did not deny existing services to any participants, as it evaluated a new service that Harambee was considering adding to an existing program. The service involved a time cost for both participants and the provider, and existing research did not clearly show if the service was likely to work. Participants in the study were not exposed to additional risks beyond their existing job search activities.

Results and policy lessons

Providing LinkedIn training to young, low-income job seekers increased end-of-program employment, with effects persisting twelve months after the program ended. 

Training and employment: The LinkedIn training increased participant end-of-program employment rates by 7 percentage points from the comparison group average of 70 percent (a 10 percent increase). This effect persisted twelve months later. The training program also increased weekly hours worked after six and twelve months. Researchers estimated that program participants saw increases in earnings of US$420 to US$1,100, using assumptions about wages from sector averages. This was 8 to 23 times higher than the per-participant cost of the intervention.      

Training and LinkedIn use: The provision of LinkedIn training increased the share of participants with LinkedIn accounts by 31 percentage points from the comparison group average of 48 percent (a 65 percent increase). Training also increased an index measuring LinkedIn usage after training completion. This suggests that program participants were not only more likely to open a LinkedIn account during the training, but also to use the account after training completion.

Relating LinkedIn use and employment effects: There is some evidence that suggests the positive impacts on employment were explained by job seekers using LinkedIn to acquire information about prospective employers. Program participants viewed more on-platform job advertisements and profiles of other users, suggesting that the LinkedIn training reduced information barriers by helping job seekers learn more about potential jobs or employers. There is mixed evidence, however, that improving information for employers or providing employee referrals via LinkedIn led to positive employment outcomes for job seekers. 

Taken together, these findings demonstrate how training young job seekers to join and use an online networking platform, LinkedIn, improves labor market outcomes and may reduce information barriers in the labor market in a relatively low-cost manner. 

Harambee has used these research findings to encourage its program participants to use LinkedIn during their job search. The study results have also informed Harambee’s development of its own job search and matching platform, SAYouth.mobi.
 

Wheeler, Laurel, Robert Garlick, Eric Johnson, Patrick Shaw, and Marissa Gargano. (2022). “Linked(to) Job Opportunities: Experimental Evidence from Job Readiness Training.” American Economic Journal: Applied Economics 14, no. 2, (April): 101-25. DOI: 10.1257/app.20200025

1.