Does Skill Certification Improve Labor Market Outcomes for Vulnerable Workers? Experimental Evidence from a Public Labor Skill Certification Program in Peru
How to cite
Marcano, Jesus; Maldonado, Stanislao; Maris, Lesbia. 2022. "Does Skill Certification Improve Labor Market Outcomes for Vulnerable Workers? Experimental Evidence from a Public Labor Skill Certification Program in Peru." AEA RCT Registry. AEARCTR-0008151. https://doi.org/10.1257/rct.8151-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
This study aims to estimate the causal effects of a public skill certification program on employment and wages of vulnerable workers in Peru. Considering the existence of low levels of employment and wages in poor and low skilled population, due to the presence of information costs on market participants, adverse selection situations on the side of workers and employers, and collective action problems in employment and wage decisions, the Peruvian Ministry of Labor and Employment Promotion (MTPE) has launched the National Program for the Promotion of Labor Opportunities "Impulsa Perú," a skill certification program to promote the achievement of quality jobs in highly vulnerable segments of the Peruvian population. Despite the fact that the program has managed to certify more than 12,000 people since 2011, its effectiveness in achieving the proposed objectives has not been studied. Additionally, it has been identified that these types of programs have not been frequently studied in the literature. Therefore, the evaluation of the causal effects of this program has the potential to shed light on the usefulness of this type of initiatives to promote quality employment in developing countries.
Experimental design
All candidates in the target population were randomly assigned to an experimental condition using a sequential block-randomized design at the city level (Imbens and Rubin 2015; Gerber and Green 2012). In this design, candidates were preselected by CCLs personnel in groups of twenty people and randomized into treatment and control conditions within the group by the research team. This strategy was determined by how the program operates: the program’s recruiters visit firms, syndicates, and related institutions to identify and recruit potential beneficiaries. The research team agreed with the implementing partner to use a sequential oversubscription strategy and recruit at least twice the available slots for the certification process, so half of the candidates can be randomized into the treatment at the group level. This strategy was sequentially implemented in 33 groups of 20 people each.
The research team randomly selected candidates into treatment after stratifying on the firm or institution where candidates were preselected using the re-randomization algorithm by Morgan and Rubin (2012). This algorithm avoids the risk of pre-treatment imbalance for a given set of covariates by allowing treatment re-randomization without affecting the design’s statistical properties. We allowed up to 1,000 re-randomizations and a minimum p-value of 0.3 to accept a given randomization. We chose to stratify in this way because the implementing partner wanted to ensure that all participating firms or organizations had employees or members in the group of beneficiaries.
Design details
Cited by
No citation data available for this registration yet.