Tobias Gabel Christiansen defends his PhD thesis at the Department of Economics

Candidate:

Tobias Gabel Christiansen, Department of Economics, University of Copenhagen

Title:

Empirical Essays on Tax Enforcement and Compliance

Supervisors:

  • Andreas Bjerre-Nielsen, Associate Professor, Department of Economics, University of Copenhagen
  • Claus Thustrup Kreiner, Professor, Department of Economics, University of Copenhagen

Assessment Committee:

  • Niels Johannesen, Professor, Department of Economics, University of Copenhagen
  • James Alm, Emeritus Professor, Department of Economics, Tulane University
  • Vedran Sekara, Associate Professor, IT University of Copenhagen

Summary:

This thesis consists of three self-contained chapters. In the first chapter, I study the long-run effects of randomized tax audits and show that taxpayers’ behavioral responses depend critically on their intentions: unintentional non-compliers substantially increase future compliance - generating revenue gains more than three times the initial audit finding. In contrast, those who deliberately evade taxes show no increase in compliance. This finding suggests that tax agencies can optimize their resources and achieve better compliance by targeting those who respond strongly to audits or by adopting more cost-effective measures such as personalized guidance to reduce inadvertent misreporting. 

In the second chapter, written together with Johanne Søndergaard, we show that self-employed individuals and principal shareholders of SMEs use person-to-person mobile payment platforms to conceal income. Announcing access to transaction data has small effects on compliance unless combined with audits, which significantly reduce transfers among non-compliant taxpayers. We further show how real-time digital traces can strengthen tax enforcement when used in conjunction with audits.

In the third chapter, written together with Christian Igel and Andreas Bjerre-Nielsen, we evaluate a large-scale randomized audit experiment at the Danish Tax Agency and find that machine-learning–guided case selection improves case-worker productivity and increases net revenue relative to current practices.

An electronic copy of the dissertation can be requested here: lema@econ.ku.dk