Essays on the political economy of development

• Main Author: Montenegro, Mateo,Ph. D.(Juan Mateo Montenegro Zarama)Massachusetts Institute of Technology.
• Other Authors: Daron Acemoglu and Esther Duflo.
• Format: Others
• Language: English
• Published: Massachusetts Institute of Technology 2020
• Subjects: Economics.
• Online Access: https://hdl.handle.net/1721.1/127032

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Economics, May, 2020 === Cataloged from the official PDF of thesis. === Includes bibliographical references (pages 157-169). === This thesis consists of three chapters which address different questions about the political economy of development. In the first chapter, Natala Garbiras-Díaz and I study whether crowdsourcing technologies aimed at augmenting civil oversight of elections might increase electoral integrity. We report the results of two large-scale field experiments we designed to assess the effectiveness of online crowdsourcing technologies in increasing the engagement of civil society in electoral monitoring around elections in Colombia. In these experiments, we leveraged Facebook advertisements to encourage citizen reporting of electoral irregularities through official websites, and also varied whether candidates were informed about the campaign in a subset of municipalities. We find that these interventions had effects on two different margins. === In addition to the expected informational effects - whereby citizen reports increased, and politicians reduced their engagement in electoral irregularities - the results highlight powerful salience effects, which operated by making electoral irregularities more top-of-mind to citizens. Specifically, the advertisements generated a large shift in the vote share of candidates perceived to be less corrupt and away from those perceived to be more corrupt. We argue that these salience effects are driven by a shift in voter preferences towards candidates they perceived as 'cleaner'. We formally test this hypothesis in a second, follow-up experiment in which we vary the salience of electoral irregularities in the advertisements sent through Facebook. As expected, we find that the advertisements featuring messages emphasizing the salience of electoral misdeeds generate a larger shift in the votes for 'cleaner' candidates than the ones only providing information about the reporting website. === The second chapter provides evidence on enforcement spillovers across enforcement activities. In particular, it shows that public audits, aimed at detecting and sanctioning corruption by public servants, increase tax compliance in Brazil. As a source of identification, it uses the geographic and time variation induced by a large-scale random audit program conducted by Brazilian federal government on municipal governments throughout the 2003-2015 period. I begin by showing that municipalities receiving an audit in the past experience an increase in federal, but not municipal tax collection. I show evidence that these effects operate through a state capacity signaling channel, whereby audits and the subsequent penal actions, act as signals both of the capacity and the willingness of the federal government to enforce the law in general, which induces citizens to increase tax compliance. === Consistent with this interpretation I show that local information about the audits, such as the one conveyed through local media or to neighboring municipalities, is key in determining the magnitude of these spillover effects across types of enforcement. The third chapter studies whether more decentralized public auditing institutions are better at increasing government accountability and reducing corruption than centralized ones. To answer this question I exploit the exogenous variation in the level of decentralization of local auditing institutions created by Colombian law to implement a regression discontinuity design and study the empirical effects of decentralizing public auditing. Using data from third-party investigations on corruption, I find that more centralized auditors do a better job at curbing corruption than decentralized ones. This result is driven by types of corruption related to public procurement as well as 'influence peddling'. === Furthermore, I find that 'effort' of public auditing institutions do not change with respect to whether these institutions are decentralized or not, which validates the use of the third-party investigations about corruption as a measure that does not confound the efforts of auditing institutions. Finally, I show evidence suggesting that the rules governing the appointment of decentralized auditors is an important mechanism in explaining the results in this setting. === by Mateo Montenegro. === Ph. D. === Ph.D. Massachusetts Institute of Technology, Department of Economics