To attend the conference, please contact the secretariat Elvire Jalran
- Helmuth CREMER, UT1, Toulouse School of Economics, Directeur de thèse
- Lucie GADENNE, Warwick University, Rapporteure
- Eduardo SOUZA-RODRIGUES, University of Toronto, Rapporteur
- Stefan AMBEC, Toulouse School of Economics, Co-directeur de thèse
- Mathias REYNAERT, UT1, Toulouse School of Economics, Examinateur
- Hélène OLLIVIER, Paris School of Economics, Examinatrice
Chapter 1: Effectively fighting deforestation requires monitoring of vast areas, which is possible thanks to satellite imagery. However, satellite monitoring can only reduce deforestation if three conditions are met: the monitoring alerts must be informative, the enforcement agency must use them to target inspections, and farmers must respond to enforcement action by doing less deforestation. This paper quantifies the contribution of real-time monitoring in deforestation reduction using detailed satellite and administrative data in the Brazilian Amazon forest. It studies the whole chain of events from the production of a deforestation alert to its effect on deforestation. It first documents an improvement in the monitoring system's ability to detect infractions in real-time. Then it estimates the impact that real-time alerts have on deforestation inspections. Finally, it estimates the impact of inspections on deforestation using an instrumental variable approach and an event study. Overall, the real-time alerts increase by three percentage points the inspection probability for offenders, avoiding approximately 450 square kilometers of deforestation per year.
Chapter 2: Developing economies are characterized by limited compliance with government regulation, such as taxation. Resources for enforcement are scarce and audit cases are often selected in a discretionary manner. We study whether the increasing availability of digitized data help improve audit targeting. Leveraging a field experiment at scale in Senegal, we compare tax audits selected by inspectors to audits selected by a risk-scoring algorithm. We find that inspector-selected audits are more likely to be conducted, to uncover tax evasion and to detect larger amounts of evasion. We show, however, that the tax administration invests less manpower in algorithm-selected cases, and that algorithm-selected audits may generate less corruption, based on survey results. In ongoing work, we attempt to unpack the algorithm’s (dis)functioning and the relevance of human capital in the audit selection and implementation process.
Chapter 3: Tax evasion is a nuisance for governments, which must devote resources to fight it to ensure that taxpayers pay their taxes. However, if taxpayers invest avoided taxes in a productive way, governments can also benefit from evasion by taxing the outcome of taxpayers’ investments. Moreover, by auditing past tax declarations, governments can still recover avoided taxes from the past while still benefiting from the result of past evasion. This amounts to a form of double taxation. This paper models tax evasion by firms in a dynamic setting where firms have incentives to invest all their assets. It shows that the optimal policy for the government is not to reduce evasion to zero, even when all enforcement parameters are free. In practice, evasion functions as a loan from the government to the taxpayer, where expected fines work as interest rates. The incentives outlined in this paper are likely to hold for small, financially constrained firms with high growth potential.