Using CAPTCHA and honesty primes to increase tax collection in Guatemala

How to cite

Sanders, Michael; Kettle, Stewart. 2014. "Using CAPTCHA and honesty primes to increase tax collection in Guatemala." AEA RCT Registry. AEARCTR-0000424. https://doi.org/10.1257/rct.424-1.0

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

Michael Sanders* , Stewart Kettle

* Primary investigator

Registry ID
AEARCTR-0000424
DOI
https://doi.org/10.1257/rct.424-1.0
First registered
Last updated
Status
In development
Trial dates
2014-07-27 – 2015-03-31
Countries
Private
Secondary IDs
TP4014019
Registry record
http://www.socialscienceregistry.org/trials/424

Abstract

This trial involves prompting and priming honesty among Guatemalan taxpayers filing declarations for income tax (ISR) or value added tax (IVA). Messages designed to induce more honest declarations are included as part of a series of CAPTCHA security systems which participants must complete in order to complete their declaration forms for these taxes.
The primary objective of the trial is to increase honesty in tax declaration. Following previous research in this area, we argue that more honest declarations will typically be report a higher tax liability than dishonest ones. The trial is important as Guatemala has the lowest rate of tax collection in Latin America (OECD, 2014).
The trial involves all people eligible to pay at least one of four taxes who complete a tax return online between July and November 2014 (approximately 1,600,000 individuals), who will be randomly allocated to either see one of six honesty prompts, or be part of a control group. Unlike other studies of tax compliance, this study targets people who are in the process of declaring their tax, rather than late or non-payers.
The intervention is delivered at the point of accessing an online declaration form. Participants are required to complete a CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) that verifies that they are a human being completing the form rather than an automated process. Our interventions consist of a series of messages inserted into the CAPTCHA pop-up.
The trial aims to test the differential impact of the original CAPTCHA (which contains no behavioural prompts), and various prompts inspired by behavioural science. These include an honesty declaration, information about public goods paid for by taxes and punishment for noncompliance, and allowing participants to choose what they believe to be a good use for public funds, or an appropriate punishment. We hypothesise that these primes will induce more honest (higher) tax declarations through various cognitive channels.

Intervention

The trial involves seven arms; a control arm where taxpayers receive the original CAPTCHA, and six adapted versions. The original Spanish versions of the treatment CAPTCHAs are included in appendix A. The CAPTCHA appears after the taxpayer selects the tax form to fill in on the main Declaraguate website, and before the form page.

Experimental design

We adopt a randomised controlled trial design with parallel conditions and individual level randomisation. Participants (Guatemalans declaring their tax liability), will be randomly assigned to see one of 7 CAPTCHAs (control or one of the six treatment arms) after selecting the tax form to fill in on the main Declaraguate website, and before the form page.
This trial will be conducted over the course of 4 months. There are four types of tax included in this trial described below . The trial involves approximately 4% of Guatemalans, and all Guatemalans who declare online for these four tax regimes. For income tax and VAT general this is 100% of declarations as online declaration is mandatory for these tax regimes, for VAT small taxpayers approximately 90% of declarations are made online.

Design details

Randomization method
Randomised by javascript randomisation code
Randomization unit
Individual
Clusters
1,600,000
Observations
1,600,000
Arms
Approximately 240,000
Primary outcomes
The objective of this trial can be framed in one of two ways – to increase tax revenues in Guatemala, and to encourage more honesty in the declaration of tax liabilities. For the purposes of this trial we take the two measures to be equivalent. Following Shu et al (2012), we assume that if a participant is reporting their tax liability dishonestly (as opposed to making an error), they will systematically understate their liability relative to the truth. We further assume that misreporting through error is statistically similar to classical measurement error, and so is randomly distributed around the true value according to some unknown distribution. For the purposes of our power calculations, we assume that this error is (a) classical (b) orthogonal on treatment, and (c) does not interact with treatment – i.e. that the measurement error does not effect treatment. If (a) does not hold, we may experience a loss of power, and any treatment effect detected would be a lower bound on the true effect size; (b) holds by design of the trial, and the validity of (c) is irrelevant if (b) holds. Hence, total tax liability declared is our primary outcome measure.

Cited by

No citation data available for this registration yet.

Keywords