The Effect of Matching on Worker Skills versus Preferences on Firm-Worker Matches in South Africa
- Job seekers
- Recruitment and hiring
Low-income countries’ labor markets are often inefficient. Workers and firms face relatively long search durations, low-quality matches between firms’ needs and job seekers’ skills, and high worker turnover. Researchers conducted a randomized evaluation to improve the quality of job matches in South Africa. They matched applicants to entry-level vacancies using different types of information: workers’ preferences for job types; communication and socio-emotional skills; or grades and work experience.
Many African countries report slow employment growth and low labor productivity in manufacturing and service jobs. This may be due to information frictions in formal sector labor markets. That is, when hiring, firms have limited information about workers’ skills and workers’ valuation of jobs. Information frictions can lead to poorly-matched hires, raising worker turnover through quits and firings/layoffs and potentially reducing productivity. Previous research has suggested that firms are willing to pay for information about workers’ skills, and that workers who can send credible signals of their skills or past performance are more likely to be employed. 1 What types of information can help improve firm-worker matches? Furthermore, what are the effects of providing new and different information on turnover, productivity, and firms’ demand for laborers?
Context of the evaluation
South Africa has one of the highest unemployment rates in the world, reaching 34.4 percent in 2021. 2 More than 60 percent of those unemployed are between the ages of 18-35. 3 Women are also more likely to be unemployed than men. 4 Despite high unemployment rates, turnover in the labor market is high. Previous evidence shows that employers in South Africa are willing to pay for more information about job seekers, and job seekers that were able to signal their skills and previous experience were more successful in finding employment. 56 This suggests a potential benefit for increasing information flows between employers and job seekers in the labor market.
In the context of the Covid-induced recession, interventions to facilitate faster firm-worker matching may be particularly important. South Africa has a relatively large non-agricultural formal sector, which was largely shut down at the beginning of the Covid-19 pandemic. By the end of 2020, 1.5 million full-time-equivalent workers (11 percent of the formal workforce) had registered job losses. 7 With this in mind, the Harambee Youth Employment Accelerator, a South Africa-based social enterprise, works to address the barriers to high-quality firm-worker matches. Harambee assesses the skills of young, inexperienced job seekers from low-income backgrounds and matches them to entry-level vacancies in the 500 client firms they work with.
Details of the intervention
Researchers partnered with the Harambee Youth Employment Accelerator to evaluate the impact of matching firms with job seekers based on either their skills or preferences. They selected over 1000 vacancies at multiple Harambee client firms in multiple sectors to take part in the study. The branches were randomly assigned to receive one of the following types of candidate shortlists:
• Skills-based shortlist: Shortlisted candidates were ranked based on different measures of soft skills. This shortlist method aimed to address potential mismatches of firms’ knowledge of a candidate’s soft skills, which can lead to layoffs or firings, as well as increase access to job opportunities for candidates from disadvantaged backgrounds with less access to formal education.
• Qualification-based shortlist: Candidates in this group were shortlisted and ranked using “business as usual” measures of grades and work experience.
• Mixed shortlist: Candidates in this group were ranked using combinations of soft skills and qualifications.
Researchers collected data from 3,600 job seekers and 1,800 managers directly after hiring decisions were made, as well as follow-up surveys 12 months later, to measure firings/layoffs, quits, promotions, performance evaluations, earnings and job satisfaction. They also collected payroll and productivity administrative data from the firms. The skills measures used in this study have been co-developed with local partners; piloted and validated; and tested against inclusivity benchmarks before being used in matching research.
Results and policy lessons
Research ongoing; results forthcoming.