Methods to Reanalyze Tax Compliance Experiments: Monte Carlo Simulations and Decision Time Analysis

Abstract

Tax compliance experiments are widely used in public finance and economic psychology. To analyze the data of the experiments, standard econometric and psychometric techniques are used. In this article, the authors show that it may be useful to employ additional data analysis tools to gain better statistical confidence on the results and to retrieve more information from the data sets, respectively. To do this, the authors reanalyze data from the tax compliance experiments of Kastlunger et al. These experiments provide evidence that after tax audits, the rates of tax compliance decrease systematically (‘‘bomb crater effect’’). First, the statistical validity of this result is tested using Monte Carlo simulations. Second, the authors extend the analysis of Kastlunger et al. using by-products of the data that had not been used in the original article to find out whether the tax compliance decisions were taken automatically or thoughtfully.