Can algorithms enhance bureaucrats’ work in developing countries? In data-poor environments, bureaucrats often exercise discretion over key decisions, such as audit selection. Exploiting newly digitized micro-data, this study conducted an at-scale field experiment whereby half of Senegal’s annual audit program was selected by tax inspectors and the other half by a transparent risk-scoring algorithm. The algorithm-selected audits were 17 percentage points less likely to be conducted, detected 47% less evasion, were less cost-effective, and did not reduce corruption. Moreover, even a machine-learning algorithm would only have moderately raised detected evasion. These results are consistent with bureaucrats’ expertise, the task complexity, and inherent data limitations.
| Repository name | URI |
|---|---|
| Reproducible Research Repository (World Bank) | https://reproducibility.worldbank.org |
Paper exhibits were reproduced on a computer with the following specifications:
• OS: Windows 11 Enterprise
• Processor: Intel(R) Core(TM) i5-1145G7 CPU @ 2.60GHz
• Memory available: 15.7 GB
Runtime: ~5 hours.
To reproduce the findings in this paper, users should follow the steps below.
Acquire the restricted data and place it in the appropriate folders. More information is provided in the data entry section. In addition, the package includes a file describing the expected folder structure for the restricted data, for users who are able to obtain access (READ ME scripts and datasets).
Run the code in the intended order after updating the global main path in each script. The scripts should be executed as follows:
Master Tax do.MASTER Prepare Analysis do. MASTER Analysis do.Construct_Data_RF_For_All_Logs do. 00_Main R. Senegal_Lit_Graphdo. Running these scripts will generate all figures and tables presented in the final paper.
Because most of the data is restricted and users may not be able to run the full workflow, the results verified by the replicators are included in the Outputs folder. This allows users to verify that the results correspond to the findings reported in the paper.
Almost all data is restricted and has not been included in the reproducibility package. For more details, please refer to the README file.
| Author | Affiliation | |
|---|---|---|
| Anne Brockmeyer | World Bank | abrockmeyer@worldbank.org |
| Pierre Bachas | World Bank | pbachas@worldbank.org |
| Alipio Ferreira | Southern Methodist University | alipioferreira@smu.edu |
| Bassirou Sarr | Senegal Ministry of Finance | basarr@minfinances.sn |
2026-01-07
| Location | Code |
|---|---|
| Senegal | SEN |
The materials in the reproducibility packages are distributed as they were prepared by the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this event do not necessarily reflect the views of the World Bank, the Executive Directors of the World Bank, or the governments they represent. The World Bank does not guarantee the accuracy of the materials included in the reproducibility package.
| Name | URI |
|---|---|
| Modified BSD3 | https://opensource.org/license/bsd-3-clause/ |
| Name | Affiliation | |
|---|---|---|
| Anne Brockmeyer | World Bank | abrockmeyer@worldbank.org |
| Reproducibility WBG | World Bank | reproducibility@worldbank.org |
| Name | Abbreviation | Affiliation | Role |
|---|---|---|---|
| Reproducibility WBG | DECDI | World Bank - Development Impact Department | Verification and preparation of metadata |
2026-01-07
1