Forecasts at scale: General equilibrium impacts of climate adaptation

How to cite

Burlig, Fiona; Jina, Amir; Kelley, Erin; Lane, Gregory. 2024. "Forecasts at scale: General equilibrium impacts of climate adaptation." AEA RCT Registry. AEARCTR-0013590. https://doi.org/10.1257/rct.13590-3.0

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

Fiona Burlig* , Amir Jina , Erin Kelley , Gregory Lane

* Primary investigator

Registry ID
AEARCTR-0013590
DOI
https://doi.org/10.1257/rct.13590-3.0
First registered
Last updated
Status
In development
Trial dates
2024-05-19 – 2026-06-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/13590

Abstract

Climate change will increase weather risk for small-holder farmers in low- income countries. We propose a novel approach to addressing this risk: accurate, long-range forecasts, which provide farmers information well in advance of the growing season about seasonal weather realizations. This project builds on Burlig et al (2024), a successful RCT in 250 villages which demonstrated that forecasts of the timing of the Indian summer monsoon's onset led farmers to substantially adjust up-front investments such as land use, crop choice, and inputs. In this project, we partner with the Indian Ministry of Agriculture and Farmers' Welfare (MoA\&FW) to deliver monsoon onset forecasts to millions of farmers across Telangana, using a low-cost SMS platform operated by the government. We embed a randomized saturation design into this scale-up, enabling us to measure general equilibrium impacts of these forecasts, including information transfer between farmers and impacts on crop prices.

Intervention

We will send SMS messages containing information about the timing of the onset of the 2024 Indian summer monsoon to farmers in a randomized saturation design, randomized at the mandal (subdistrict) level.

Experimental design

We use a randomized saturation design at the mandal level, randomizing (while stratifying on district) 172 mandals to "high-saturation" (90% of farmers on the SMS extension platform to receive messages) and 171 manuals to "low-saturation" (30% of farmers on the SMS platform to receive messages). We will then randomize message delivery at the individual level according to these saturation levels, stratified by mandal.

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
Mandal level for saturation, individual level for messages
Clusters
343
Observations
Approximately 3,000 individuals in our survey sample
Arms
172 mandals high-saturation, 171 mandals low-saturation, ~800k farmers treatment (no message), ~1.2m farmers control (message).
Primary outcomes
monsoon onset beliefs (this version of the PAP)

Cited by

No citation data available for this registration yet.

Keywords