Family Resource Allocation and Tax Decisions

How to cite

Chen, Kuan-Ming. 2025. "Family Resource Allocation and Tax Decisions." AEA RCT Registry. AEARCTR-0015787. https://doi.org/10.1257/rct.15787-1.0

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

Kuan-Ming Chen*

* Primary investigator

Registry ID
AEARCTR-0015787
DOI
https://doi.org/10.1257/rct.15787-1.0
First registered
Last updated
Status
In development
Trial dates
2025-04-14 – 2025-12-31
Countries
Private
JEL codes
G5, H2, D1
Registry record
http://www.socialscienceregistry.org/trials/15787

Abstract

In the Taiwanese context, the tax system allows older individuals to be claimed as dependents by a designated family member—typically one of their adult children. When claimed as a dependent, the adult child receives a deduction from their taxable income. If a parent has multiple adult children, it is most tax-efficient for the family to designate the child with the highest income as the claimant, since that child faces the highest marginal tax rate. However, administrative data reveal that many families do not make this optimal choice. Our randomized controlled trial (RCT) is designed to identify the underlying causes of this phenomenon.

Intervention

We will conduct online surveys both before and after the tax filing month (May). The survey has three versions. The control version collects only demographic information and basic tax filing details. Treatment 1 adds an informational component: respondents receive a brief quiz to assess their understanding of the tax code and learn how selecting the optimal dependent can reduce their tax liability. Treatment 2 builds on Treatment 1 by including an additional behavioral nudge, encouraging respondents to inquire about their siblings' income and consider whether it would be more tax-efficient to switch the dependency designation. After the tax filing month, we will conduct another trial to collect the filing behavior after our treatments.

Experimental design

We have one control and two treatments. We use the National Taiwan University Websurvey platform to recruit our respondents and then randomly assign them into the three groups.

Design details

Randomization method
Randomization done by the computer powered by Qualtrics.
Randomization unit
Individual
Clusters
4800 individuals.
Observations
4800 individuals.
Arms
1600 control, 1600 treatment 1, and 1600 treatment 2.
Primary outcomes
Tax filing behavior before and after the trial.

Cited by

No citation data available for this registration yet.

Keywords