We construct a new linked survey-administrative dataset for Italy to show the relevance of heterogeneity among the self-employed for income evasion. We find that liberal professionals, sport and art performers underreport to tax authorities about 50% of their true income, which is almost three times the size of income underreporting detected on average for entrepreneurs. We also find that low-educated entrepreneurs evade more income than high-educated ones and provide explanation for this result. Then, we study the implications of evasion on income dis-tribution and earning differentials. We find that households at the 10th income decile generate about 80% of tax revenue losses due to self-employment income underreporting. We also show that the self-employment earning puzzle is reversed after accounting for evasion, with income gains of about 40% relative to salaried workers with differences across working status and edu-cational level.

We present empirical evidence on the cost of tax sheltering among high-income taxpayers within progressive income tax systems. Exploring a unique personal income tax policy change in Australia, we use the “bunching” around the top tax kink to estimate the costs of tax sheltering and the Elasticity of Taxable Income (ETI). Our findings reveal substantial behavioural responses to tax changes among high-income taxpayers, particularly those with greater flexibility in income adjustment, suggesting that tax-sheltering behaviours play a pivotal role in their responses. When accounting for these costs, the estimated ETI significantly increases, emphasising the necessity of considering tax sheltering in tax policy analysis. We contribute to the ongoing discourse on optimal tax policy design and its impact on economic behaviour. Our findings have important implications for policy debates on whether high-income individuals should be taxed at higher rates.

We examine the use of paid tax preparers by individual taxpayers in the US between 2011 and 2019. We provide novel descriptive evidence characterizing the paid tax preparers and the users of their services. We show that over 55% of individuals use paid tax prepares and taxpayers who use paid tax preparers earn more than non-users, but claim significantly more credits at the same time. The average size of the tax preparer clientele varies with average income of clients and declines for clients with highest incomes, suggesting small specialized tax preparer networks for the richest clients. Second, we quantify the tax savings that paid preparers are able to offer their clients using a two-way fixed effect switchers model (AKM). Preliminary evidence suggests that while individual fixed effects determine the overwhelming majority of variation in effective income tax rates (ETRs), variation in paid preparer skill explains 10% of ETR inequality.

This paper documents the impact of third-party enforcers on behavioral tax responses. Using administrative data from the wealthiest state in Brazil, we analyze the effect of a tax reform that designated notary officers as enforcers of gift tax liability. Our findings indicate a 3.8(1.3)-fold increase in the number (total sum) of reported gifts per semester, positively impacting tax liabilities. The reform led to smaller individual gifts, with the elasticity of taxable gifts close to 0.01, which is significant after the reform. These results suggest that the delegation policies, such as pre-auditing schemes, promote tax compliance in developing countries.