A Machine Learning Approach to Analyze and Support Anti-Corruption Policy

34 Pages Posted: 28 May 2020 Last revised: 1 Jun 2020

See all articles by Elliott Ash

Elliott Ash

ETH Zürich

Sergio Galletta

University of Bergamo; ETH Zürich

Tommaso Giommoni

ETH Zürich

Date Written: April 1, 2020

Abstract

Can machine learning support better governance? In the context of Brazilian municipalities, 2001-2012, we have access to detailed accounts of local budgets and audit data on the associated fiscal corruption. Using the budget variables as predictors, we train a tree-based gradient-boosted classifier to predict the presence of corruption in held-out test data. The trained model, when applied to new data, provides a prediction-based measure of corruption which can be used for new empirical analysis or to support policy responses. We validate the empirical usefulness of this measure by replicating, and extending, some previous empirical evidence on corruption issues in Brazil. We then explore how the predictions can be used to support policies toward corruption. Our policy simulations show that, relative to the status quo policy of random audits, a targeted policy guided by the machine predictions could detect more than twice as many corrupt municipalities for the same audit rate.

Keywords: machine learning, corruption, audits, local public finance

JEL Classification: D73, E62, K14, K42

Suggested Citation

Ash, Elliott and Galletta, Sergio and Giommoni, Tommaso, A Machine Learning Approach to Analyze and Support Anti-Corruption Policy (April 1, 2020). Available at SSRN: https://ssrn.com/abstract=3589545 or http://dx.doi.org/10.2139/ssrn.3589545

Elliott Ash (Contact Author)

ETH Zürich ( email )

Rämistrasse 101
ZUE F7
Zürich, 8092
Switzerland

Sergio Galletta

University of Bergamo ( email )

Via Salvecchio, 19
Bergamo, 24129
Italy

ETH Zürich ( email )

Rämistrasse 101
ZUE F7
Zürich, 8092
Switzerland

Tommaso Giommoni

ETH Zürich ( email )

Rämistrasse 101
ZUE F7
Zürich, 8092
Switzerland

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
221
Abstract Views
988
rank
161,178
PlumX Metrics