Can Public Employment Programmes Reduce Long-term Unemployment?

How to cite

Calland, Jack; Franklin, Simon; Garlick, Rob; Orkin, Kate. 2026. "Can Public Employment Programmes Reduce Long-term Unemployment?." AEA RCT Registry. AEARCTR-0019300. https://doi.org/10.1257/rct.19300-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Jack Calland* , Simon Franklin , Rob Garlick , Kate Orkin

* Primary investigator

Registry ID
AEARCTR-0019300
DOI
https://doi.org/10.1257/rct.19300-1.0
First registered
Last updated
Status
On going
Trial dates
2026-05-01 – 2027-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/19300

Abstract

Public works programmes (PWPs) are one of the most popular forms of anti-poverty policy around the world. There is growing recognition of their short-run / contemporaneous benefits in rural settings, but evidence of longer-term / post-programme effects in urban settings is scarce. We study the City of Cape Town's Expanded Public Works Programme (EPWP), which randomly selects beneficiaries for short-term work opportunities. Surveying beneficiaries with programme experience up to over ten years ago, we assess whether the programme improves long-term outcomes, including employment, earnings, skills and mental health, and how these vary over job characteristics and over recipient characteristics, including gender.

Intervention

See the "Experimental design" section.

Experimental design

We exploit the randomized allocation of public works jobs to unemployed jobseekers to study the causal effects of the program on employment and mental health outcomes. Since 2013, the City of Cape Town has allocated EPWP opportunities using a centralized algorithm that randomly selects workers from a database of unemployed jobseekers. Project managers at the City set up projects and hire workers from the database. For each work opportunity on a project, the manager can define basic eligibility criteria, such as address, minimum educational qualification, or disability status. Jobseekers are randomly selected from the (usually larger) pool of jobseekers satisfying these criteria, who are then returned on a randomly-ranked waitlist. Managers then call workers on the list to offer them work on the project along with relevant details (pay, duration, location, etc.) The randomized waitlist structure means that different individuals have different probabilities of being called with work offers, which offer causal analysis of the treatment effects of the program.

Design details

Randomization method
The City of Cape Town uses a centralized algorithm to conduct randomization. The algorithm is a SQL stored procedure that randomly selects and ranks jobseekers from the relevant tables of the database.
Randomization unit
Individual
Clusters
Not applicable.
Observations
We select 19,316 individuals waitlisted for 1,969 opportunities out of a possible 630,818 individuals and 37,220 opportunities on the database. The database contains a very large number of participants but not all are suitable for analysis due to missing data, timing of work, etc. The number we can contact depends on the cost and time needed to track people, which we will only learn during the surveys. We expect to successfully reach a subset of the 19,316 sampled individuals. If funds and timeline allow, we may sample more individuals from the database.
Arms
9,658 received EPWP work; 9,658 did not receive EPWP work
Primary outcomes
Our primary outcomes are current employment and earnings.
Secondary outcomes
Our other outcomes are all mechanisms that link the treatment to the primary outcomes and/or welfare measures. These include other job attributes besides earnings (contract type, occupation, employer size, commute time, cost and mode, job satisfaction), employment history both in EPWP and outside EPWP, job search, job market expectations, mental health, risky behavior, consumption, and related topics.

Cited by

No citation data available for this registration yet.

Keywords