Targeting Payments for Ecosystem Services by Risk in Meghalaya, India

Last registered on April 03, 2025
View Trial History

Pre-Trial

Trial Information

General Information

Title
Targeting Payments for Ecosystem Services by Risk in Meghalaya, India
RCT ID
AEARCTR-0015675
Initial registration date
March 30, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
April 03, 2025, 1:04 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country
India
Region
Meghalaya

Primary Investigator

Name
Lucy Page
Affiliation
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Maike Pfeiffer
PI Affiliation
Yale University
Contact Investigator
PI Name
Rohini Pande
PI Affiliation
Yale University
Contact Investigator

Additional Trial Information

Status
On going
Start date
2025-01-27
End date
2029-01-01
Keywords
Environment & Energy
Additional Keywords
Payments for ecosystem services, Machine learning, Targeting
JEL code(s)
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Protecting ecosystems in low-income countries is crucial to climate-change mitigation and adaptation. Payments for ecosystem services (PES)—which compensate landowners for refraining from ecologically damaging activities—are a promising tool to protect ecosystems while providing income to rural communities. Central to designing effective PES programs is ensuring that they yield “additionality,” or conservation that would not have occurred without payments. Voluntary enrollment, and thus self-selection, into PES programs tends to limit additionality: those who would conserve their land even in the absence of the program have the highest incentive to enroll. Past work suggests that machine-learning models can use basic remote-sensing data to predict deforestation risk at a granular spatial scale. Then, targeting PES enrollment by predicted deforestation risk may be an effective tool to increase additionality. In this randomized control trial, we collaborate with the Government of Meghalaya, India, to embed machine-learning predictions of deforestation risk into a large-scale expansion of an existing PES program. In particular, we will test whether targeting higher-payment PES contracts to landowners on land with higher predicted risk of deforestation can improve program additionality and cost-effectiveness. We will use a combination of satellite, administrative, and primary data to test the impacts of risk-based targeting on PES enrollment, deforestation, program cost-effectiveness, and household welfare.
External Link(s)

Registration Citation

Citation
Page, Lucy, Rohini Pande and Maike Pfeiffer. 2025. "Targeting Payments for Ecosystem Services by Risk in Meghalaya, India." AEA RCT Registry. April 03. https://doi.org/10.1257/rct.15675-1.0
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
Our primary intervention is targeting PES contracts by predicted risk of deforestation, which we predict at a 30m spatial scale using a convolutional neural network trained on satellite data layers. We classify land into a binary of high- vs. low-risk land, and the government offers higher-payment PES contracts to high-risk land.
Intervention Start Date
2025-01-27
Intervention End Date
2029-01-01

Primary Outcomes

Primary Outcomes (end points)
Forest loss, carbon storage, and enrollment payments by predicted risk, allowing us to assess the cost effectiveness of averted deforestation across PES treatment arms.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We randomize villages between three groups:

Basic PES: PES program staff will approach villages to actively recruit applicants for PES. Government representatives will invite anyone who owns forest land with area over 1ha to enroll in contracts under which they commit to refraining from damaging activities for 30 years, including cutting trees, mining, grazing animals, and raising structures. Enrollees will receive payments at a fixed per-hectare rate for the first 5 years of the contract. These payments will be set at Rs. 10,000 per hectare, with additional payments offered for land where trees are particularly dense or are otherwise of particular ecosystem value because they are e.g. close to a national park or fall in an elephant corridor. Government staff will assess landowners' compliance with PES contracts each year and will reduce enrollees' next-year payments if they have been found in violation.

Risk-targeted PES: In villages enrolled in this arm, the government will offer higher PES payments on land at higher predicted risk of deforestation; otherwise, PES recruitment remains the same as in the base PES program. We predict risk of deforestation at a 30m scale using a convolutional neural network trained on satellite data. In particular, landowners in these villages will be offered payments of Rs. 16,000 per hectare per year for each unit of high-risk land they enroll in Green PES, while being offered Rs. 8,000 per hectare per year for low-risk land. Payments are calibrated such that the average PES payment offered in this arm is equal to that offered in basic PES, as we classify about 25% of land in risk-targeted villages as high-risk.

Control group: No active PES recruitment will occur in control villages. If landowners in control villages learn about Green PES and contact program staff to register, they will be allowed to enroll. However, program staff will not actively visit these villages.
Experimental Design Details
Not available
Randomization Method
Randomization done via Stata.
Randomization Unit
Village-level randomization
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
581 villages
Sample size: planned number of observations
581 villages
Sample size (or number of clusters) by treatment arms
Basic PES: 194 villages

Risk-targeted PES: 193 villages

Control: 194 villages
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
IFMR Human Subjects Committee
IRB Approval Date
2024-12-23
IRB Approval Number
IFMRIRB-IEIC-1224-2-A
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information