Dataset for the paper "Who’ll stop lying under oath? Empirical evidence from tax evasion games" (2020)

[1] : Centre d'Économie de la Sorbonne, University Paris 1 Panthéon-Sorbonne (UMR8174)
[2] : Paris School of Economics
[3] : Aix-Marseille School of Economics, Aix-Marseille Université (UMR7316)
[4] : Centre de recherche sur les entreprises (Burgundy School of Business)
[5] : Department of Economics, College of Business, University of Wyoming
Description :
Dataset from two earned income/tax declaration experimental designs used to address the question of who actually responds to an oath with honesty, and why.
Discipline :
economics (social sciences)

General metadata

Data acquisition date : from 2014 to 2016
Data acquisition methods :
  • Experimental data :
    The oath procedure implemented in the experiment closely follows the design of Jacquemet, Joule, Luchini, and Shogren (2013). After filling out the consent form, the monitor asks each subject to sign a solemn oath form
    which states that he or she “swear upon [his/her] honor that, during the whole experiment, [he/she] will tell the truth and provide honest answers” (in bold in the original form).

    Experiment 1 followed a two step design—earned income and declared income. First, to strengthen external validity, the experiment starts with a real effort task in which subjects earn their income. The earned-income task is similar to Alm, Cherry, Jones, and McKee (2012): subjects sort numbers in ascending order in a 3 * 3 matrix filled with digits generated in random order. We compute the earnings based on the time taken to complete the task - the quicker a subject is, the more money s/he earns, according to the following compensation rule (labeled in an Experimental Currency Unit): 150 − (time ⇥ 13). The task is repeated 5 times and earned income is the sum of earnings from all 5 tasks. Second, subjects are asked to “declare the amount of income they earned at the previous stage”. They can declare any amount from no income up to 100% of earned income. We recorded the time taken by each subject to state their income declaration. The tax rate in the experiment is 35% and the money declared through taxes is to be donated to the World Wide Fund for Nature (WWF). Donations are certified directly by theWWF, with certificates sent directly to the participants by email. Last, subjects are presented with debriefing questionnaires to collect their socio-demographics.Each session lasted about 1 hour, with an average payoff of 20 euros (17 euros directly given to the participants and 3 euros donated to WWF), including a 5 euro show-up fee.

    Experiment 2 is identical to Experiment 1 with one exception - in step two, we repeat the declaration stage five times. We make only one decision binding by telling subjects that we will select one declaration at random to determine their net income (i.e., their experimental earnings) and their donation to WWF. The decision context is kept constant and the sequence elicits the subjects’ degree of behavioral indecision. Our subject pool is again divided into two treatments: baseline and oath.
Language : English (eng)
Formats : application/x-stata-dta, text/csv
Audience : Research
Publications :

DOI and links

10.25666/DATAUBFC-2025-01-24
https://dx.doi.org/doi:10.25666/DATAUBFC-2025-01-24
https://search-data.ubfc.fr/FR-13002091000019-2025-01-24

Quotation

Nicolas Jacquemet, Stéphane Luchini, Antoine Malézieux, Jason F. Shogren (2020): Dataset for the paper "Who’ll stop lying under oath? Empirical evidence from tax evasion games". BSB. doi:10.25666/DATAUBFC-2025-01-24

Record created 24 Jan 2025 by Cécile Schweitzer.
Local identifier: FR-13002091000019-2025-01-24.

dat@BSB

dat@BSB is a sub-portal of dat@UBFC, a metadata catalogue for research data produced at UBFC.

République Française
dat@UBFC
dat@BSB
Université de Bourgogne, Université de Franche-Comté, UTBM, AgroSup Dijon, ENSMM, BSB, Arts des Metiers