Davi Bhering

This paper analyzes how capital disclosure, repatriation, and allocation change in response to increased offshore monitoring. It leverages Brazil's 2016 offshore tax amnesty, implemented ahead of the expansion in monitoring capacity under the global Automatic Exchange of Information framework. Using administrative microdata on the universe of reported foreign assets, we find that 21,000 individuals disclosed $59 billion (3.4% of GDP) in the amnesty. Participation was strongly correlated with wealth, reaching 21% among the top 0.001%. However, net capital repatriation was limited, averaging only 20% of disclosed assets. Most wealth remained abroad, driven by high foreign returns and diversification benefits. Tracking repatriated capital, we find it flowed mainly into domestic fixed income rather than the real economy. Consequently, firms owned by amnesty participants showed no growth in employment or revenues, although they significantly substituted external debt with proprietary capital. These results suggest that while amnesties effectively uncover hidden wealth, they are insufficient to fundamentally shift capital allocation.

We study the redistributive and efficiency effects of differentiated commodity taxation in a constrained tax system. Using administrative data and a large VAT cut in India, we show that tax reductions are only partially passed through to prices, generate modest quantity responses, and induce little product mislabeling. As a result, well-targeted differentiated commodity taxes can achieve substantial redistribution at moderate efficiency cost. Our calibrations yield an optimal luxury tax premium of 20 percentage points. Given the features of India's income tax and transfer system, we show that overturning this conclusion would require large and simultaneous relaxation along all margins.

We combine population-wide administrative micro-data to provide new estimates of income inequality and effective tax rates by income groups in Brazil, capturing all income and all tax payments. Our data allow us to link businesses to their owners and thus to allocate business income and associated taxes to the corresponding individual firm owners. The top 1% earns 27.4% of total income in 2019, one of the highest levels recorded in the world. The tax system, which relies heavily on consumption taxes, is regressive: while the average tax rate in the economy is 42.5%, this rate falls to 20.6% for million-dollar earners, due to the non-taxation of dividends and provisions that reduce corporate tax liabilities.

This paper constructs new estimates of income distribution across Brazilian states from 2006 to 2019, combining survey, tax, and national accounts data within the Distributional National Accounts framework. This approach improves on survey-based measures by correcting the underrepresentation of top incomes, incorporating income components typically missing from surveys and tax records, and aligning the results with macroeconomic aggregates. The estimates reveal substantial regional differences in inequality that are not fully captured by conventional survey data, particularly because the importance of capital income varies across states. Amazonas, Rio de Janeiro, and São Paulo emerge as the most unequal states, with the top 1% receiving around 28% of total pre-tax income and the top 10% nearly 60%, while Amapá, Acre, Rondônia, and Santa Catarina show the lowest concentration levels. The results also suggest that income concentration remained broadly stable across states over the period analyzed. By providing harmonized state-level estimates consistent with national income, the paper offers a more accurate and internationally comparable picture of regional inequality in Brazil.