Behavioural Nudges and Income Bunching below Tax Thresholds: Evidence from Randomised Controlled Trials in Indonesia

How to cite

Satyadini, Agung; Effendi, Subagio; Pradikta, Rheno; Tantripal, Leonard; Otlani, Rizmy. 2026. "Behavioural Nudges and Income Bunching below Tax Thresholds: Evidence from Randomised Controlled Trials in Indonesia." AEA RCT Registry. AEARCTR-0017769. https://doi.org/10.1257/rct.17769-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Agung Satyadini* , Subagio Effendi , Rheno Pradikta , Leonard Tantripal , Rizmy Otlani

* Primary investigator

Registry ID
AEARCTR-0017769
DOI
https://doi.org/10.1257/rct.17769-1.0
First registered
Last updated
Status
On going
Trial dates
2025-12-15 – 2026-04-30
Countries
Indonesia (All Provinces)
JEL codes
C93, H25, H26
Registry record
http://www.socialscienceregistry.org/trials/17769

Abstract

The aim of this trial is to investigate how behvioural nudges affect income reporting behaviour around statutory tax threshold, with a particular focus on income bunching among individual taxpayers in Indonesia. Using a large-scale randomised controlled trials, the intervention delivers targeted informational (norms and public goods provision), and deterrence-based emails to around 850,000 taxpayers whose reported income lies close to the threshold that triggers higher tax obligations. The analysis evaluates whether these nudges influence both the likelihood and magnitude of income bunching below the threshold, and whether observed responses persist over time.

Intervention

The intervention consists of behaviourally informed email communications sent to individual taxpayers whose reported income lies within a predefined bandwidth around a statutory tax threshold that triggers higher tax obligations. The objective of the intervention is to influence income reporting behaviour, particularly the propensity to report income just below the threshold.

Eligible taxpayers are randomly assigned to one of three treatment arms, each receiving a single email conveying a distinct behavioural message:

Deterrence email: This message emphasises statutory filing obligations and reporting deadlines, the administrative penalty for late or incorrect filing. It also highlights the possibility of further enforcement actions in cases of non-compliance

Localised peers norms email: This message provides information on income reporting behaviour among geographically proximate or economically similar taxpayers. Specifically, it highlights the proportion of comparable taxpayers in the recipient’s local area or peer group who report income above the statutory threshold, with the aim of conveying prevailing compliance norms.

Public goods provision email: This message highlights the role of individual income tax contributions in financing public expenditure, with particular emphasis on government spending related to education.

All messages are delivered through official Directorate General of Taxes email communication channels. Visual elements, including colours and graphical layout, are designed to enhance message salience and readability, drawing on established findings in the behavioural and colour psychology literature.

Experimental design

The study employs a large-scale randomised controlled trial to examine how behavioural nudges affect income reporting behaviour around statutory tax thresholds. The experimental sample consists of individual taxpayers whose baseline reported income lies within a predefined bandwidth around the threshold that triggers higher tax obligations. Eligible taxpayers are randomly assigned at the individual level to one of several treatment arms—deterrence, public goods provisions, or localised peer norms—or to a control group that receives no additional communication. Randomisation is stratified on pre-treatment characteristics, including baseline income, distance to the threshold, compliance history, sector, gender, and geographic location, to ensure balance across groups. All interventions are delivered through official tax authority email channels during the same intervention window, allowing for a clean comparison of treatment effects on subsequent income reporting and bunching behaviour over time

Design details

Randomization method
The study employs an individual-level stratified randomisation design. Eligible taxpayers are first partitioned into strata based on pre-treatment characteristics, including baseline reported income bins around the statutory tax threshold, distance to the threshold, prior compliance history, economic sector, gender, and geographic location. Within each stratum, taxpayers are randomly assigned to one of the treatment arms—deterrence, tax literacy, or localised peer norms—or to a control group using a computer-generated random number process. The randomisation is implemented using a fixed random seed to ensure reproducibility, and assignment probabilities are identical across strata. This procedure ensures balance in observable baseline characteristics across experimental groups while preserving the internal validity of the causal estimates.
Randomization unit
individual
Clusters
-
Observations
850,000 taxpayers
Arms
control: 77,700 T1: 230,000 T2:230,000 T3:230,000
Primary outcomes
Reported taxable income relative to the threshold. Reported taxable income compared to the previous period.
Secondary outcomes
Total tax liability reported. Filing timeliness. Subsequent compliance in later tax periods

Cited by

No citation data available for this registration yet.

Keywords