Behavioral interventions and tax compliance: evidence from a field experiment

Last registered on September 21, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
Behavioral interventions and tax compliance: evidence from a field experiment
RCT ID
AEARCTR-0019585
Initial registration date
September 14, 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
September 21, 2026, 9:25 AM EDT

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
Lucas Rosenzvit
Affiliation
Universidad de San Andrés
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Juana María Sojo
PI Affiliation
Universidad de San Andrés
Contact Investigator

Additional Trial Information

Status
On going
Start date
2026-09-01
End date
2026-10-20
Keywords
Behavior, Finance & Microfinance, Governance
Additional Keywords
Tax-compliance, debt
JEL code(s)
H26
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
This thesis examines whether an extraordinary municipal debt communication increases tax compliance and whether adding information about the use of public resources produces an additional effect. We implement a randomized controlled field experiment involving more than 16.000 municipal accounts. Accounts are assigned with equal probability, within 48 strata, to four arms: no debt statement; a standard debt statement without an additional municipal message; a statement containing a public-safety message; or a statement containing a public-space message. Stratification is based on valid-email availability, property locality, and four categories of overdue installments.
Accounts in the three active treatment arms with a valid email address receive the same statement through both physical and digital channels.

The primary outcome is full repayment of the outstanding debt through payment option 1, using either the QR code or barcode, within a 30-day follow-up period. The main secondary outcomes are enrollment in a payment plan through option 2, payment of current bill 20265A/B, and payment of the full debt in 1 installment without using the QR code number 1. The primary analysis follows an intention-to-treat approach and uses account-level regressions with randomization-stratum fixed effects and heteroskedasticity-robust standard errors. The design identifies the effect of receiving a debt statement, the incremental effect of adding a public-service message, and the difference between the two messages. At the time of writing, outcome data are not yet available.
External Link(s)

Registration Citation

Citation
Rosenzvit, Lucas and Juana María Sojo. 2026. "Behavioral interventions and tax compliance: evidence from a field experiment." AEA RCT Registry. September 21. https://doi.org/10.1257/rct.19585-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
The study evaluates whether sending a municipal debt statement, and whether adding different messages to that statement, affects debt repayment. Eligible municipal accounts are randomly assigned with equal probability to one of four groups: no debt statement; a debt statement without an additional municipal message; the same debt statement with a public-safety message; or the same debt statement with a public-space message. When an account assigned to receive a statement has a valid email address on file, the same statement is also sent by email. Email availability is a pre-treatment account characteristic and is not randomized.
Intervention Start Date
2026-09-04
Intervention End Date
2026-10-09

Primary Outcomes

Primary Outcomes (end points)
Full payment of the total outstanding debt shown in the debt statement through payment option 1, recorded within the 30-day follow-up window beginning on September 3, 2026.
Primary Outcomes (explanation)
The primary outcome is a binary account-level indicator equal to one when the administrative records show full payment of the total debt through option 1 during the specified 30-day window, and zero otherwise. Payments may be initiated through the QR code or the barcode printed on the statement. The outcome will be measured for all randomized accounts according to their original assignment, following an intention-to-treat approach.

Secondary Outcomes

Secondary Outcomes (end points)
1. Enrollment in a payment plan through payment option 2 within the 30-day follow-up window.
2. Payment of the current municipal bill, identified as bill 20265A/B, within the 30-day follow-up window.
3. Payment of the total outstanding debt using othen methods than the QR code n1 option.
Secondary Outcomes (explanation)
Each secondary outcome will be constructed as an account-level binary indicator using municipal administrative records. The first equals one when the account enrolls in a payment plan through option 2 during the follow-up period. The second equals one when payment of bill 20265A/B is recorded during that period. Both outcomes will be analyzed by original randomized assignment.

Experimental Design

Experimental Design
This is a four-arm, account-level randomized controlled trial using administrative data. Eligible accounts are assigned with equal probability to a no-statement control group or to one of three debt-statement groups. Randomization is stratified by valid-email availability, property locality, and four approximately equal-sized categories of the number of overdue installments. The primary analysis will compare outcomes across assigned groups using an intention-to-treat framework.
Experimental Design Details
Not available
Randomization Method
Computer-generated stratified randomization using a fixed random seed (3676586226). The algorithm assigned each eligible account within each of 48 pre-defined strata to C0, B, S, or O with a target probability of 0.25 per arm. When a stratum size was not divisible by four, the remainder was allocated so that realized arm counts within that stratum differed by no more than one account. The seed and the final input database were preserved to make the assignment exactly reproducible.
Randomization Unit
Individual municipal account.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
Not applicable: the treatment is not clustered. There are 16,125 individual account-level randomization units.
Sample size: planned number of observations
16,125 municipal accounts in the randomized experimental sample.
Sample size (or number of clusters) by treatment arms
C0 — no debt statement: 4,031 accounts.
B — debt statement without an additional municipal message: 4,031 accounts.
S — debt statement with the public-safety message: 4,032 accounts.
O — debt statement with the public-space message: 4,031 accounts.
Total randomized sample: 16,125 accounts.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
The study uses the full eligible administrative population available at the time of randomization; the sample size was therefore not chosen on the basis of an ex ante power calculation. No minimum detectable effect size was used to determine enrollment. Any reported MDE will be calculated transparently from the realized arm sizes and explicitly stated assumptions, rather than presented as a design input.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
CEDLAS Research Ethics Committee
IRB Approval Date
2026-08-27
IRB Approval Number
27082026
Analysis Plan