Since the late 1970s, many countries have based their tax systems on self-assessment – taxpayers are expected to evaluate their liabilities autonomously, and voluntarily remit their tax due. If the tax system is perceived as fair and easy to navigate, with credible threat of penalisation for non-compliance, self-assessment reduces the cost of tax administration without significant revenue losses (Barr et al. 1977; Teviotdale and Thompson 1999; James and Alley 2004). On the other hand, self-assessment entails an increase in compliance costs for taxpayers, at the very least in terms of time spent complying with their obligations. However, none of the conditions mentioned above – fairness, simplicity and credibility – is easy to meet. Hence, initial moves towards self-assessment were met in many countries with an increased focus on what type of deterrence measures would increase taxpayer compliance (Forest and Sheffrin 2002), following the prevalent theoretical approach of the time (Allingham and Sandmo 1972). By the late 1990s, the focus was shifting to the perceived fairness and complexity of the tax system, increasingly seen as both a direct and indirect obstacle to compliance (Slemrod and Venkatesh 2002; Forest and Sheffrin 2002; Eichfelder and Schorn 2012). Intuitively, a taxpayer who does not understand their tax obligations has a hard time complying with them, and might well decide not to try at all – especially if penalisation is seen as unlikely.