Research at J-PAL Africa
The J-PAL Africa research team conducts randomised evaluations on social programmes and policies in South Africa. Topics covered to date include labour markets, education, urban services, health and political participation.
We have worked on the following evaluations with South African researchers:
Job Counselling, Productivity Signals and Employment
Researchers Martin Abel, Rulof Burger and Patrizio Piraino are aiming to identify efficient devices to credibly signal the productivity of young unemployed to employers in South Africa. Their work includes testing the impact of a job search counselling program, the use of a reference letter template and personal action plans on job search behaviour and employment outcomes of young job seekers. The study is a collaborative effort with the South African Department of Labour to assist the government in reducing youth unemployment via a rigorous impact evaluation of alternative employment services.
Overcoming information barriers to firms hiring young, inexperienced work-seekers in South Africa
Researchers Eliana Carranza, Robert Garlick, Kate Orkin, and Neil Rankin are running two concurrent studies to evaluate information barriers in the labour market for work-seekers and for small and medium enterprises. The studies examine i) whether providing SMEs with information on work-seekers' abilities can lead to improved hiring decisions, and ii) providing work-seekers with information on their own abilities will improve their chance of employment and/or better orient their strategies for finding employment. The study runs in partnership with Harambee, an organisation that screens and evaluates young, entry-level work-seekers and connects them to employers looking for entry level talent.
Information on Contestation and Turnout in South Africa's First Competitive Election
Political participation in elections is considered a core component of a vibrant democracy and a key mechanism for strengthening accountability of political leaders. Yet the paradox of voting suggests that voters have little incentive to participate since the probability their individual vote will impact the election outcome is close to zero. Kate Orkin conducted this lab-in-the-field study in South Africa’s 2016 municipal elections. It tests the effect of different messages aiming to shift voters’ calculus and encourage the perception that their votes do count. Around 2,000 registered voters, recruited from around Soweto in Johannesburg, were randomly assigned to receive different messages or be in a control group that was simply surveyed. South Africa uses a mixed member system in municipal elections, so citizens’ votes count both in their ward but also directly for which party wins and governs the city. Some treatment group voters simply received face-to-face messages giving better information on this municipal voting process. Other groups were additionally informed that the elections would be closely contested. Using self-reported and behavioural data, the experiment examines the effect of the different messages on beliefs, turnout and political engagement.
Researchers Kelsey Jack and Grant Smith partner with the City of Cape Town to evaluate the impact of switching from a credit-based billing model to prepaid meters on household electricity consumption, welfare, and utility company revenues. For more detailed information on the study, read their policy brief here.