Analysis of Behavioural Interventions' Spillover Effects on Individual Taxpayers Compliance: Evidence from Two Randomised Trials in Indonesia
How to cite
Yogama, Eko Arief; Rablen, Matthew D.; Gray, Daniel J.. 2024. "Analysis of Behavioural Interventions' Spillover Effects on Individual Taxpayers Compliance: Evidence from Two Randomised Trials in Indonesia." AEA RCT Registry. AEARCTR-0013600. https://doi.org/10.1257/rct.13600-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
Behavioural interventions often create unintended effects on the other aspects of tax compliance that were not targeted during the design stage, or known as spillover effects. This study utilises data collected from two randomised controlled experiments that have been done in Indonesia to investigate the spillover effects that may arise from the interventions to annual income tax return compliance. In the previous experiments, we examined the impact of the interventions involving individual taxpayers. The sample consist of more than 51,000 taxpayer, where half of the sample were allocated to Control (status quo-received no treatment) and the other half to Treatment Group (received either physical letters or digital messages as the intervention). To investigate whether there are spillover effects from those two previous experiments, an instrumental variable regressions (IV) were used. The endogenous variables are instrumented by the initial random allocations to represent assuming that the taxpayers who although allocated to treatment group but in reality did not received the intervention would not be impacted by the treatments. This study could be a reference for policymakers when designing interventions. When designed should use the results as a reference when designing compliance-related interventions so it could cover a broader spectrum of tax compliance.
Intervention
We performed two field experiments consisting of just above 10,000 individual taxpayers in the first trial and around 41,000 individual taxpayers in the second trial. In both experiments, the sample is equally divided into Control and Treatment groups.
Experimental design
In the previous trials, the sample consist of just above 10,000 individual taxpayers in the first trial and around 41,000 individual taxpayers in the second trial. In both experiments, the sample is equally divided into Control and Treatment groups. The treatment in the first trial is sending physical letters accompanying a standard penalty notification form. The treatment in the second trial is digital messages broadcasted through WhatsApp reminding of the obligation to pay monthly income taxes.
Design details
Cited by
No citation data available for this registration yet.