Do Nudges Improve Firms' Compliance with Environmental Regulations?
Last registered on August 09, 2018


Trial Information
General Information
Do Nudges Improve Firms' Compliance with Environmental Regulations?
Initial registration date
August 07, 2018
Last updated
August 09, 2018 1:33 AM EDT

This section is unavailable to the public. Use the button below to request access to this information.

Request Information
Primary Investigator
University of Chicago
Other Primary Investigator(s)
PI Affiliation
University of Chicago
PI Affiliation
University of Chicago
Additional Trial Information
In development
Start date
End date
Secondary IDs
The goal of this project is to reduce effluent violations under the Clean Water Act (CWA) National Pollutant Discharge Elimination System (NPDES) permit program. The U.S. Environmental Protection Agency (EPA) has developed an automated, facility-targeted messaging system to encourage compliance with NPDES permit limits and reduce overall effluent levels. This study will examine how effective early forms of notifications or warnings could be in inducing facilities to self-correct problems before they become chronic violators. We will evaluate the impact of this notification system on compliance through a randomized controlled trial (RCT), which will generate evidence on both the overall effectiveness of the system at changing effluent violation rates and which configurations of messaging duration and recipient have the greatest impact on facility effluent discharges.
External Link(s)
Registration Citation
Armstrong, Sarah, Ludovica Gazze and Michael Greenstone. 2018. "Do Nudges Improve Firms' Compliance with Environmental Regulations?." AEA RCT Registry. August 09.
Experimental Details
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
* Discharge index: The (quarterly) mean standardized effluent level of the ith company at time t

* Discharge relative to permit: The (quarterly) mean standardized effluent level relative to the permitted level of the ith company at time t

* Probability of violation A compliance indicator, equal to one (1) if company i reported at least one effluent limit exceedance during time t and 0 otherwise

We standardize effluent levels within groups defined by various Discharge Monitoring Report and permit limit value characteristics, including unit of measurement and whether a value is a concentration or quantity. This results in a standardized (z-score) measure of effluent levels, and effluent levels as standardized deviations relative to the permit limit. Because our randomization is implemented at the company-level, our primary outcome variables are company-level indices (means) of these standardized effluent levels at the quarter-year level.
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
The aim of this experiment is to determine whether automatically generated notifications incentivize compliance with the Clean Water Act and reduce overall effluent levels. Companies in four U.S. states were randomly assigned to one of four treatment groups that vary by the duration of time in which a company is eligible to receive notifications and whether upper management within the company receive notifications in addition to the employee that submitted the data detailing waste water discharges.

Companies are randomly assigned to one of the four following treatment groups: (1) Companies are eligible to receive notifications upon each effluent exceedance identification throughout the study; (2) companies are eligible to receive notifications for a shortened duration during the study; (3) companies are eligible to receive notifications throughout the study and signatories' managers also receive notifications; or, (4) companies receive no notifications. Additionally, half of facilities associated with companies assigned to (1) or (2) are assigned to receive no notifications during the study.

Experimental Design Details
Not available
Randomization Method
We randomly assign companies to treatment groups using a minimum-maximum p-value re-randomization algorithm (Bruhn & McKenzie, 2009) to improve balance of key variables across treatment groups (balance covariates).
Randomization Unit
Our primary unit of randomization is parent companies, with half of facilities with parent companies assigned to groups (1) and (2) assigned to receive no intervention.
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
370 parent companies
Sample size: planned number of observations
438 facilities
Sample size (or number of clusters) by treatment arms
Treatment group (1) 114 parent companies; treatment group (2) 120 parent companies; treatment group (3) 29 parent companies; treatment group (4) 107 companies (control group)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
Social and Behavioral Sciences Institutional Review Board
IRB Approval Date
IRB Approval Number