Does providing information on government spending alter taxpayer beliefs and compliance? Evidence from a large-scale field trial in Belgium

This field trial explores the potential impact of introducing a pie-chart with relative spending across the major categories of public goods and services at the beginning of the tax reporting process. This experiment is run on the universe of taxpayers who file their taxes online in Belgium. Randomization happens by way of the last two digits of the national identity number. The treatment group is divided into four sub-groups that test for the additional impact of introducing a social norm message, a loss-aversion frame to the provision of public goods and tax compliance, and making penalties for non-compliance explicit. To measure compliance we consider reported taxable income and itemized tax deductions. Taxpayers are also invited to respond to a survey at the end of the tax reporting process that gauges (1) taxpayer satisfaction; (2) taxpayer appreciation of public goods and services; (3) tax fairness; (4) tax morale; (5) belief about own understanding of tax spending categories; (6) detailed beliefs about actual tax spending categories; and their (7) tax spending preferences by way of a priority ranking of spending categories. Heterogeneity analyses using the survey responses are limited to gender, year of birth, and the type of information treatment that was received. Individuals in the control group only get to see the information pie-chart after having been invited to respond to the survey. This design allows us to explore the impact of the information treatment on taxpayer beliefs as well as potential mechanisms in case the information treatment has an effect on tax compliance. This experiment ties in with a previously registered trial that tested for the impact of introducing nudges in the context of reminder letters for late filing as well as late tax payments.