Essays on public finance in developing countries

This dissertation comprises three essays on public finance in developing countries. Chapter 2 studies the real economic effects of a crackdown on offshore financial secrecy in Colombia. Exploiting variation in the probability of detecting offshore wealth triggered by the Panama Papers leak in 2016, this chapter finds that firms owned by users of offshore vehicles increase retained earnings and use this additional liquidity to increase capital and employment, resulting in higher profits. These positive real responses also lead to higher corporate income tax revenues, complementing the revenue gains from increased wealth tax compliance. Overall, this chapter provides new evidence on the real economic implications of offshore tax evasion by wealthy individuals. Chapter 3, co-authored with Daniel Guzmán, studies how value-added taxes (VAT) distort firms’ make-or-buy decisions. Using a unique input-output dataset from Chile and variation from a VAT reform, this chapter documents a partial pass-through of higher VAT into prices net of taxes of professional services. As a response to higher outsourcing costs, firms reduce purchases of professional services and substitute toward in-house employment. Despite this reallocation, we find no effects on firm growth or efficiency. These findings highlight how VAT can distort firms’ production organization, challenging the traditional view of VAT neutrality under full input tax crediting. Chapter 4, co-authored with Mateo Uribe, studies the relationship between inequality and tax collection. As opposed to models that emphasize redistribution, this chapter focuses on taxation as a mechanism for financing public goods that foster market development and increase productivity. To do so, we build a model of property tax compliance in which landowners’ decisions depend on the capitalization of public spending into land values and expected enforcement when evading taxes. We assume that both the benefits of public spending and the expected enforcement increase with land inequality. We use rich administrative data from Colombian municipalities between 1923 and 1960 to validate the model and show that its predictions are consistent with the observed data.