Nudging Taxpayers Via WhatsApp to Submit Their Annual Tax Return: Evidence From a Field Experiment in Indonesia

Last registered on February 21, 2023


Trial Information

General Information

Nudging Taxpayers Via WhatsApp to Submit Their Annual Tax Return: Evidence From a Field Experiment in Indonesia
Initial registration date
February 17, 2023

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
February 21, 2023, 10:26 AM EST

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



Primary Investigator

Erasmus University Rotterdam

Other Primary Investigator(s)

PI Affiliation
Erasmus University Rotterdam

Additional Trial Information

In development
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
The Indonesian government usually sends a reminder message via email to individual taxpayers to submit their tax returns on time. Still individual taxpayers often fail to submit their tax return or submit late. The country works with a self-assessment system. And tax authority’s standard message relies on a traditional, normative approach (reminding taxpayers about the reporting deadline). This study tests alternative messaging and delivering messages via WhatsApp in a field experiment in 5 Cities (Jakarta, Ciamis, Cilacap, Jayapura and Merauke) in Indonesia. The primary outcome is the timely submission of the annual tax return. Individuals are randomized into four groups:(1) Standard message, (2) WhatsApp - Standard message, (3) WhatsApp - Deterrence message, and (4) WhatsApp - Reciprocity message.
External Link(s)

Registration Citation

Rieger, Matthias and Rio Widianto. 2023. "Nudging Taxpayers Via WhatsApp to Submit Their Annual Tax Return: Evidence From a Field Experiment in Indonesia." AEA RCT Registry. February 21.
Experimental Details


There are four experimental conditions: (1) Control - Standard Message (2) WhatsApp - Standard Message (3) Deterrence-WhatsApp Message (4) Reciprocity-WhatsApp Message

Condition (1) receives the standard email. Condition (2) gets the same standard email but an additional standard message by WhatsApp. Condition (3) receives the standard email and a deterrence WhatsApp message. Condition (4) receives the standard email and a reciprocity WhatsApp message.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Tax return submitted
Primary Outcomes (explanation)
Tax return equals 1 if reported on time and zero otherwise (unsubmitted or submitted too late). We expect tax return rates to be higher in the treated groups (conditions 2,3,4) compared to the control/standard message (condition 1). We do not have a hypothesis which of the treated groups (conditions 2,3,4) has the largest impact.

Secondary Outcomes

Secondary Outcomes (end points)
Amount of tax payment
Secondary Outcomes (explanation)
The variable is in rupiahs and available if the taxpayer submitted a tax return. We expect the tax amounts reported to be higher in conditions 3 and 4 (compared to conditions 1 and 2) due to the reciprocity and threat content of the nudges.

Experimental Design

Experimental Design
The treatment effects will be estimated based in an intention-to-treat (ITT) model with and without covariates. Both linear probability models and probit models will be presented.
Experimental Design Details
Baseline covariates include taxpayers' characteristics (age, gender, and timely tax return status in the previous year). We will also show results with tax office dummies included.
We will present a heterogeneity analysis based on our baseline model without controls via sample splits by gender and at the sample median age.
We also intend to explore potential mechanisms relating to taxpayer decisions. A post-experimental interview will be used to collect information on taxpayers' expectations. The interview aims to answer whether the treatment mechanism works and how taxpayers respond to the treatments and delivery methods used by reporting their 2022 Annual Income Tax Return on time.
Randomization Method
Randomization is carried out in the pre-experimental stage (before sending the message). We randomly assign (intent-to-treat) the cleansed data of the individual self-employed taxpayers to four experimental conditions. We do so, balancing on age, gender, and annual tax return reporting on-time for the 2021 fiscal year. We use the randomize command in STATA. We block randomization within tax office.
Randomization Unit
individual self-employed taxpayers
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
Six tax offices in five cities (one in Ciamis, Cilacap, Jayapura, Merauke, and two in Jakarta) in four Provinces in Indonesia are included in the experiment. However, the randomization is done at the individual level. Randomization is done via blocking by tax office to ensure balanced group sizes per office (block randomization).
Sample size: planned number of observations
The potential sample size is 24,887 individual self-employed taxpayers based on historical administrative records with phone number available.
Sample size (or number of clusters) by treatment arms
The total of 24,887 taxpayers is divided into four groups.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
The potential study population covers 24,887 individual tax payers with a contact phone number drawn from administrative data in 6 tax offices across four provinces. With four groups, power of 80%, alpha of 5%, and a control proportion of 0.4035842, we are powered for a MDE of 0.0245 on the primary outcome. [power twoproportions 0.4035842. alpha (0.05) n1(6221) n2 (6221) power (0.8)].
Supporting Documents and Materials

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

Request Information

Institutional Review Boards (IRBs)

IRB Name
Erasmus University Rotterdam
IRB Approval Date
IRB Approval Number


Post Trial Information

Study Withdrawal

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

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials