Abstract
Fiscal policies frequently determine discontinuities in the average tax rate; for example, introducing fiscal benefits or flat-rate tax for those that report an income below a threshold produces a large discontinuous jump in tax liability – a notch – at the threshold. These notched tax schemes create strong and salient incentives to bunching below the threshold, either reducing their work or increasing non-compliance behaviour. Empirical evidence consistently reports a density hole in the taxpayers' distribution above the threshold in notched fiscal systems, but less than theoretically expected (e.g., Kleven and Waseem, 2013); experimental evidence supports the existence of bunching behaviour but not to the full extent what questions the understanding of incentives in notched fiscal systems (Gibson et al., 2019). This experimental investigation aims to estimate taxpayers' responses to incentives produced by the notch in personal income taxation when evasion is possible. We aim to evaluate what might be the overall impact of a notched fiscal system on the economy in terms of labour supply and tax evasion. Four between-subject treatments with 90 subjects each are implemented in the laboratory: the earnings of a real-effort task based on sliders (Gill and Prowse, 2012) are taxed with (i) a proportional scheme without evasion possibilities, (ii) a notched scheme without evasion possibilities, (iii) a proportional scheme with evasion possibilities, (iv) a notched scheme with evasion possibilities. Preliminary results show clearly bunching in notched tax systems with and without tax evasion; bunching is stronger and average effort is lower in notched tax system without tax evasion. Strong learning of incentives sticks out: subjects falling in the dominated area progressively decrease over rounds.