Digital Finance and Human Development: Global Evidence from the Financial Possibility Frontier
Description
This replication package accompanies the study “Digital Finance and Human Development: Global Evidence from the Financial Possibility Frontier.” It contains the code and redistributable data used to construct the Digital Financial Index (DFI), State Financial Capacity (SFC) measures, descriptive statistics, main instrumental-variable estimates, local-projection IV results, and robustness analyses for a cross-country panel covering 2004–2024. The package includes step-by-step replication files and a README describing data sources, variable construction, processing steps, and the order in which the analysis should be reproduced. Data are drawn from publicly available sources including the IMF Financial Access Survey, Worldwide Governance Indicators, UNDP, the Global Macro Database, and other international databases. Some third-party datasets used in the study are not redistributed because of licensing or redistribution restrictions. For these sources, the README provides the original source information and instructions for obtaining the data directly from the respective providers.
Files
Steps to reproduce
Before running the replication files, users should first obtain the third-party datasets that are not redistributed in this repository because of licensing or redistribution restrictions. For full replication, download the giant resource-discovery data associated with Arezki et al. (2016), the GSMA Mobile Connectivity data, and the state-capacity data from O’Reilly and Murphy (2022) from their original providers. Place the cleaned/source-compatible files in the raw data folder using the filenames expected by the code: giant-res_discovery.dta, gsma.dta, and state_cap_dta2.dta. The giant-discovery data are required for the main IV analysis; the GSMA data are used for supplementary digital-readiness measures; and the O’Reilly–Murphy data are used for the augmented State Financial Capacity robustness analysis. The remaining redistributable source files are included in the replication package. The master-data code explicitly merges the O’Reilly–Murphy state-capacity variables and the giant-discovery measures into the country-year panel. Then reproduce the analysis in the following order: Update the root path in 0-master-dofile.do. Run dofile-step1-index.do to construct the baseline and alternative DFI measures. Run dofile-step-2.do to merge all data sources and create the master panel. Run dofile-step-3.do to reproduce the main IV, first-stage, SFC-regime, LP-IV, and robustness results. Run dofile-step-3b-alternative-outcomes.do for PHDI and log real GDP per capita. Run dofile-step-3c-alternative SFC.do for the augmented SFC robustness analysis. Alternatively, after adding the restricted datasets and setting the path, run 0-master-dofile.do to execute the full workflow automatically.
Institutions
- RMIT UniversityVictoria, Melbourne