Coordination Frictions and the Market for Crop Residue in India

Last registered on January 28, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
Coordination Frictions and the Market for Crop Residue in India
RCT ID
AEARCTR-0017178
Initial registration date
January 27, 2026

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
January 28, 2026, 8:00 AM EST

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

Locations

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

Request Information

Primary Investigator

Name
Sheetal Sekhri
Affiliation
Sheetal Sekhri, Associate Professor, University of Virginia
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Raymond Guiteras
PI Affiliation
NC State
Contact Investigator
PI Name
Sandeep Dixit
PI Affiliation
CIPT
Contact Investigator

Additional Trial Information

Status
On going
Start date
2025-07-15
End date
2026-12-30
Keywords
Agriculture, Environment & Energy
Additional Keywords
Crop Residue Management, Stubble Burning
JEL code(s)
Q52, Q53
Secondary IDs
KCAI GR-3027
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Crop residue burning after the rice harvest is a major source of seasonal air pollution in northern India, with harmful consequences for public health and the climate. One emerging and relatively low-cost alternative to burning is the use of baler services, which mechanically collect and remove crop residue from fields. However, despite the availability of baling technology, many farmers are unable to access these services in practice. Farmers face search and coordination costs that limit timely residue removal, while baler operators require sufficient scale and reliable scheduling to operate profitably. As a result, potentially beneficial residue removal services are often not provided when they are most needed.
This study evaluates whether reducing these coordination barriers can lower crop residue burning. We conduct a cluster-randomized controlled trial across 101 villages in Punjab, India. In treatment villages, we organize information sessions and registration drives that connect farmers with a designated baler operator and facilitate scheduling of residue pickup. No monetary incentives or price subsidies are provided. Control villages continue under existing practices, in which farmers and balers may make arrangements independently but receive no coordination support from the study team.
The primary outcome is crop residue burning, measured using satellite imagery and machine-learning methods. Secondary outcomes include adoption of baling services, alternative residue management practices, agricultural yields, and self-reported health symptoms. By testing whether lowering coordination costs can induce service provision and reduce environmentally harmful behavior, this study aims to shed light on the role of market frictions in shaping environmental outcomes in agricultural settings
External Link(s)

Registration Citation

Citation
Dixit, Sandeep, Raymond Guiteras and Sheetal Sekhri. 2026. "Coordination Frictions and the Market for Crop Residue in India." AEA RCT Registry. January 28. https://doi.org/10.1257/rct.17178-1.0
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)
This cluster-randomized trial tests whether reducing coordination costs between farmers and baler operators reduces post-harvest crop residue burning. In treatment villages, we conduct information camps and registration sessions that provide farmers with information about baling services and facilitate contact and scheduling with a designated baler operator. Following harvest, the baler arranges directly with participating farmers to collect and remove crop residue at no cost to the farmer.
Control villages receive no information sessions or coordination support. Farmers and balers in control villages are free to make arrangements independently, as under the status quo, but no facilitation is provided by the study team. The trial measures the effect of this coordination intervention on crop residue burning and related outcomes.
Intervention Start Date
2025-10-15
Intervention End Date
2025-11-30

Primary Outcomes

Primary Outcomes (end points)
Probability of burning for a plot aggregated to village level.
Crop residue burning is measured using remote sensing, combining satellite imagery and supervised machine-learning methods, supplemented by limited ground-truth spot checks.
Crop residue burning is measured using remote sensing, combining satellite imagery and supervised machine-learning methods, supplemented by limited ground-truth spot checks.
Primary Outcomes (explanation)
Crop residue burning is detected using high-resolution satellite imagery (PlanetScope and Sentinel-2) combined with a supervised machine-learning classifier following a method similar to that proposed by Walker et al. (2022). At the plot level, the algorithm produces a predicted probability that the plot was burned during the post-harvest period between paddy harvest and wheat planting. A binary burning indicator is constructed based on this prediction and aggregated to the village level for analysis.

Secondary Outcomes

Secondary Outcomes (end points)
Participation in baling services (registration, pickup, and quantities collected), which captures take-up of the service and serves as a proximate mechanism linking the intervention to reductions in burning; Diversion from other crop residue management methods (non-burning) ; Rice yield and production; Farmer knowledge and attitudes toward balers and alternative residue management; Farmer satisfaction with the free pickup service; Average time between harvest and residue pickup;


Self-reported respiratory health symptoms during the burning season.
Secondary Outcomes (explanation)
We measure diversion from other non-burning crop residue management practices to model impacts on greenhouse gas emissions. Different on-site residue management practices (e.g., incorporation, composting) generate different emissions profiles. Data on non-baling practices will be collected through household surveys and combined with emissions coefficients from the climate science literature to construct measures of greenhouse gas emissions (CO2-equivalent).
Participation, knowledge, and satisfaction outcomes are measured through household surveys. Time to pickup is computed from reported harvest and collection dates. Baler utilization is derived from administrative and survey data. Health outcomes rely on survey questions regarding respiratory symptoms during the burning season.

Experimental Design

Experimental Design
The study includes 101 villages organized into 10 clusters corresponding to catchment areas around designated balers’ collection and processing facilities. Randomization to treatment and control is conducted at the village level and stratified by cluster and, within cluster, by village size (measured by total cropped area).
Treatment villages receive information camps and registration drives designed to facilitate coordination between farmers and a designated baler operator. Control villages continue under the status quo: farmers and balers may arrange residue collection independently, but no coordination or facilitation is provided by the study team.
The primary outcome, crop residue burning, is measured using satellite-based remote sensing. Participation in baling services and other secondary outcomes are measured through household surveys and administrative data.
Experimental Design Details
Not available
Randomization Method
Randomisation uses a computer script with a fixed seed to ensure reproducibility
Randomization Unit
village
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
100
Sample size: planned number of observations
2000
Sample size (or number of clusters) by treatment arms
20 farmers will be surveyed per village
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
With 100 total villages (50 treatment and 50 control), we have 80% or greater power to detect a 15 percentage point reduction in the share of acres burned. This power calculation is based on data from Ludhania showing a mean (by village) share of acres burned of 0.6 with a standard deviation (across villages) of 0.25. This is a plausible effect size if we make the conservative assumption that 40% of farmers offered the collection service will accept it, and half of this will represent additionality, i.e., half of farmers accepting collection service would have burned in the absence of the offer, while half would have managed their stubble in other ways without burning. This takeup rate is reasonable given the nearly 60% takeup by farmers in our partner's pilots.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Heartland
IRB Approval Date
2025-04-11
IRB Approval Number
041025-1188
Analysis Plan

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

Request Information