Increasing voluntary tax compliance through outreach to clients of return preparers

Last registered on July 12, 2021

Pre-Trial

Trial Information

General Information

Title
Increasing voluntary tax compliance through outreach to clients of return preparers
RCT ID
AEARCTR-0007925
Initial registration date
July 07, 2021
Last updated
July 12, 2021, 4:13 PM EDT

Locations

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

Request Information

Primary Investigator

Affiliation
OES - GSA

Other Primary Investigator(s)

PI Affiliation
Office of Evaluation Sciences, U.S. General Services Administration
PI Affiliation
Office of Evaluation Sciences, U.S. General Services Administration
PI Affiliation
Office of Evaluation Sciences, U.S. General Services Administration
PI Affiliation
Research Applied Analytics and Statistics Division, Internal Revenue Service
PI Affiliation
Office of Evaluation Sciences, U.S. General Services Administration

Additional Trial Information

Status
On going
Start date
2021-01-14
End date
2022-01-01
Secondary IDs
Abstract
This project aims to improve the outreach and education efforts to return preparers and taxpayers implemented by the Internal Revenue Service (IRS) as part of the agency’s broader strategy to reduce errors in claiming benefits by intervening early and ensuring compliance with the law. Specifically, this project aims to build evidence on the effectiveness of the pre-filing season educational letters that the IRS sends to clients of return preparers and the effectiveness of program changes to this strategy. The Letter 6138 informs clients that their return may contain errors in claiming benefits, provides tips to selecting a return preparer, and advises the taxpayer to review the return for errors in claiming certain benefits.

The General Service Administration’s (GSA’s) Office of Evaluation Sciences (OES), and the IRS’s Research, Applied Analytics and Statistics (RAAS) and Wage and Investment (W&I) divisions collaborated on this project to measure the effects of sending the Letter 6138 to clients of return preparers on a number of outcomes, including returns filed with errors claiming certain benefits and clients’ tax filing method. Additionally, this project will examine the effects of the Letter 6138 on clients who used the same return preparer, but were not themselves sent a Letter 6138 (i.e., the indirect or network effect of the Letter 6138) and explore whether the Letter 6138 affected the returns prepared by return preparers during filing season 2021. Finally, the project will examine whether sending a modified behavioral insights (BI) Letter 6138 affects the outcomes of clients differently from sending clients the treatment-as-usual (TAU) Letter 6138.

Registration Citation

Citation
Brown, Julia et al. 2021. "Increasing voluntary tax compliance through outreach to clients of return preparers ." AEA RCT Registry. July 12. https://doi.org/10.1257/rct.7925-1.1
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2021-01-14
Intervention End Date
2021-01-31

Primary Outcomes

Primary Outcomes (end points)
We have four different primary outcomes:
● Change in Filing Method in filing season 2021 (“Change Method”): this variable is binary, and it adopts the value one if a client changed their method of filing their tax return by using a different return preparer (paid return preparer or free tax preparation services) than the return preparer they used during the 2020 filing season, submitted their own tax return, or did not file a tax return.
● Tax Benefit Error: this variable is binary, and it adopts the value one if the client files a return that may contain one or more errors in claiming certain benefits, which for the purposes of this study include: the earned income tax credit (EITC), child tax credit/additional child tax credit/credit for other dependents (CTC/ACTC/ODC), American opportunity tax credit (AOTC) and head of household (HOH) filing status. It is zero otherwise, including if the client does not claim these benefits or does not file a tax return.
● Refund Amount: this variable is numeric, and it reflects the return-level refund amount. For the purposes of this study, this is a measure that will be used to calculate protected revenue, which captures the monetary savings from sending the Letter 6138 to clients.
● Sum of erroneous dollars: this is a numeric variable that equals the numeric value for the sum of erroneous dollars that may have been claimed for certain benefits on each return. This is an alternate measure which will be used to calculate protected revenue.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
We have six additional outcomes that we consider secondary:

Earned Income Tax Credit (EITC) Error: is a binary variable that adopts the value one if the return may contain an error when claiming EITC.

American Opportunity Tax Credit (AOTC) Error: this variable is binary, and it adopts the value one if the return may contain an error when claiming AOTC.

Combined Child Tax Credit Error: this variable is binary, and it adopts the value one if the return may contain an error when claiming ACTC/CTC/ODC.

Review Return: this is a binary variable that adopts the value one if the client reviewed their return from filing season 2020 and filed a 1040X modification to correct any errors in that return..

Self-File: this is a binary variable that adopts the value one if the client filed their tax return during the 2021 filing season themselves, and 0 otherwise.

Did Not File: this is a binary variable that adopts the value one if the client did not file their tax return during the 2021 filing season.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
This project involves a two-step randomization process. In the first step of the randomization process (or the first level of randomization), we implemented a cluster-randomized design where the cluster is defined as a group of clients who used the same return preparer during filing season 2020. In the second step of the randomization process, our sample includes returns filed by the return preparer during the 2020 filing season that may contain errors in claiming certain benefits. In the second step of the randomization process (or the second level of randomization), we randomized these returns (i.e., clients) in treatment return preparer clusters to be sent the BI Letter 6138, TAU Letter 6138, or not sent any pre-filing season letter.

At the first level of randomization, we grouped together similar return preparers into return preparer blocks and randomized return preparers within these blocks. We aim for each block to have 6 return preparers; however, block size ranges from 3-8 return preparers. Within these return preparer blocks, we randomized two-thirds of return preparers to have the IRS send some clients the Letter 6138 and one third of return preparers to have no clients sent any pre-filing season letter.

At the second level of randomization, we randomize clients, whose return preparers are assigned to the treatment group, in client random assignment blocks. We grouped together clients who used the same return preparer and had similar returns during the 2020 filing season. Within each client block, we randomized approximately 40 percent of clients to be sent a Letter 6138 and approximately 60 percent of clients to be sent no pre-filing season letter. Finally, among clients randomized to be sent a pre-filing season letter, we randomized the version of the Letter 6138 the IRS sent them. Within each client block, half o]f clients sent pre-filing season letters were sent the TAU Letter 6138 and half were sent the BI Letter 6138.
Experimental Design Details
Not available
Randomization Method
Randomization done in office by a computer
Randomization Unit
In the first step of the randomization process (or the first level of randomization), we implemented a cluster-randomized design where the cluster is defined as a group of clients who used the same return preparer during filing season 2020. In the second step of the randomization process, our sample includes returns filed by the return preparer during the 2020 filing season that may contain errors in claiming certain benefits. In the second step of the randomization process (or the second level of randomization), we randomized these returns (i.e., clients) in treatment return preparer clusters to be sent the BI Letter 6138, TAU Letter 6138, or not sent any pre-filing season letter.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
There are 1994 return preparer clusters.
Sample size: planned number of observations
There are 77,692 clients/returns (unit of analysis).
Sample size (or number of clusters) by treatment arms
1,331 return preparer clusters treatment (52,348 clients: 10,425 BI letter, 10,424 TAU letter, and 31,499 no letter), 663 return preparer clusters control (25,344 clients)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Institutional Review Board, Office of Research Integrity, American University
IRB Approval Date
2021-03-03
IRB Approval Number
IRB-2021-323