Acitvating Informal Labour for Public Service Delivery: ReCycling in Mozambique
How to cite
Mendonça, Frederica. 2026. "Acitvating Informal Labour for Public Service Delivery: ReCycling in Mozambique." AEA RCT Registry. AEARCTR-0017589. https://doi.org/10.1257/rct.17589-1.0
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 paper studies whether jointly activating household behaviour and underutilised informal labour can improve last-mile service provision. I implement a randomised controlled trial in Quelimane, Mozambique, combining household information on recycling with an institutional innovation that repurposes bicycle taxi drivers as waste collectors through a municipal buyback centre. A second treatment arm adds a conspicuous signalling device for households. The paper tests whether low-cost institutional arrangements that activate informal labour networks can help overcome coordination failures and improve public service delivery in low-capacity urban settings.
Intervention
I conduct a field experiment (RCT) in which an information-based recycling intervention was delivered to randomly selected households in the city of Quelimane, Mozambique. The project was implemented in partnership with the municipal government under the name “ReCiclar com Quelimane'' (ReCycling with Quelimane). The intervention ran from July to November 2025 and consisted of a household-level information campaign on recycling, delivered through door-to-door visits, combined with free collection of recyclable waste. Recyclable waste collection was carried out by bicycle taxi drivers (“cyclists”), who constitute the city’s main transport network. Cyclists were engaged as waste collectors in a manner that complemented their regular work as bicycle taxis.
Experimental design
The allocation to treatment was conducted at the level of the cyclist. I sampled up to 3 clients per cyclist. All clients and cyclists allocated to treatment received the intervention; clients and cyclists in the control group received no intervention. No intervention was conduc Cyclists were randomly allocated to one of these groups:
- a group receiving the simple treatment (T1);
- a group receiving the nudge treatment (T2);
- a control group receiving no intervention.
Design details
Cited by
No citation data available for this registration yet.