The Determinants of Income Tax Compliance: Evidence from a Controlled Experiment in Minnesota

Abstract

This paper reports on the results of a controlled experiment in Minnesota in which a random sample of taxpayers was informed that their income tax returns would certainly be closely examined. We analyze reported income of this sample of taxpayers, reported income on their previous year's returns, and reported income from the two corresponding years' returns of a control group of taxpayers that did not receive the letter. We find that the treatment effect varies depending on the level of income. Low and middle income taxpayers increased reported income and tax liability relative to the control group, which we interpret as indicating the presence of noncompliance. The effect was much stronger for those with more opportunity' to evade, as measured by their source of income. However, the reported income of the high-income treatment group fell sharply relative to the control group. We suggest a model based on tax audits as a negotiation that can explain this apparently perverse result.