Person:

Naritomi, Joana

This dissertation comprises three chapters. The first chapter investigates whether consumers can help governments improve firm compliance with the Value Added Tax. It exploits quasi-experimental variation from a government program from Sao Paulo, Brazil that created monetary rewards for consumers to ask for receipts. To assess how incentives to consumers can be effective despite potential collusion opportunities, I construct datasets for 1 million firms, 40 million consumers, and 2.7 billion receipts. I estimate that revenue reported in retail increased by at least 22% over four years. The estimated compliance effect is stronger for sectors with a high volume of transactions and small receipt values, consistent with a model in which there are fixed costs to negotiate collusive deals to avoid issuing receipts. Furthermore, the effect has an inverted-U shape with respect to firm size. This result is consistent with a model of higher baseline compliance among larger firms, and in which shifts in audit probability from consumer monitoring increase in firm size. I find no effects on exit rates or formal employment decisions.