The Effects of Penalty Information on Tax Compliance: Evidence from a New Zealand Field Experiment

Abstract

The ‘standard’ Allingham-Sandmo-Yitzhaki (ASY) model of tax evasion predicts effects on compliance which depend on the perceived probability of detection, tax rate and penalty for evasion. Compliance effects of detection probabilities and tax rates have been extensively tested empirically, but penalty effects are rarely tested explicitly. This paper examines the effects of late payment penalties on tax compliance based on an experiment involving New Zealand goods and service tax (GST) ‘late payers’. Firstly, based on an ASY-type model of tax late payments in which the probability of enforcement, rather than detection, is central, we develop a number of testable hypotheses. Secondly, based on a field experiment involving a specific compliance intervention, we examine how taxpayers respond when given different penalty information. The experiment also allows us to consider differences between taxpayers’ stated intentions to comply and subsequently observed compliance. Results suggest that differences in penalty information given to taxpayers and reductions in penalty rates both affect taxpayers stated intentions to comply (pay overdue tax and penalties) as predicted. However, subsequently observed responses generally appear unresponsive to penalties. Nevertheless, various individual taxpayer characteristics are identifiable that affect both compliance intentions and actual behaviour.