Trading off investment in human capital versus productive assets: perceptions of risk and return
How to cite
Noritomo, Yuma; Morsink, Karlijn; Shikuku, Kelvin. 2026. "Trading off investment in human capital versus productive assets: perceptions of risk and return." AEA RCT Registry. AEARCTR-0017654. https://doi.org/10.1257/rct.17654-1.0
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
Agriculture has long been the main and often the only career path for low-income farmers in many developing countries. In many contexts, however, climate change is making agriculture increasingly risky, while new labor market opportunities for skilled workers are emerging. This study examines how the decisions of parents to invest in the education and labor of their children in pastoralist households in northern Kenya and southern Ethiopia are influenced by uncertainty about climate risks and future labor markets. Using hypothetical choice experiments and a randomized intervention that provides information on climate risk and labor market opportunities, the research investigates how expectations about the future, conditional on parental preferences and beliefs, influence investments in education and productive assets.
Intervention
Our treatment intervention is designed to inform households about climate risk and labor market conditions. We have three treatment arms: climate risk information, labor market information, and a combined information arm (see the analysis plan for further details).
(1) Households assigned to the climate risk information group receive information on climate risk, including the number of droughts in their region during the past 25 years and trends over time.
(2) Households assigned to the labor market information group receive information on labor market conditions, including average wages and employment probabilities conditional on educational attainment.
(3) Households assigned to the combined information group receive both types of information described above.
Experimental design
The proposed randomized experiment includes approximately 2,000 pastoralist households with school-aged boys. During the survey, households are randomly assigned to one of four groups, with approximately 500 households in each group:
(1) Households assigned to the climate risk information group receive information on climate risk, including the number of droughts in their region over the past 25 years and trends over time.
(2) Households assigned to the labor market information group receive information on labor market conditions, including average wages and employment probabilities conditional on educational attainment.
(3) Households assigned to the combined information group receive both types of information described above.
(4) Households in the control group do not receive any information.
Further details are provided in the pre-analysis plan.
Design details
Cited by
No citation data available for this registration yet.