Local Banking Structure and Business Use of Public Instant Payments: Evidence from Brazil's Pix - results replication package

# Brazil Pix JFSR Final Paper Replication Package This package reproduces the empirical tables, figures, online-appendix exhibits, and key output checks for the final JFSR paper on local banking structure and business use of Brazil's Pix instant-payment system. ## Contents - `code/`: Python scripts for cleaned-data and raw-data replication. - `data/cleaned/`: frozen processed data for the cleaned-data path. - `docs/`: data sources, codebook, raw upload list, and package manifest. - `reference_outputs/`: reference JFSR tables, figures, online-appendix exhibits, table/figure crosswalk, in-text numbers, and checksums. The main manuscript file is not included. The editable online appendix is generated as a replication output. ## Environment Create an environment with either pip or conda: ```bash python3 -m venv .venv source .venv/bin/activate python3 -m pip install -r requirements.txt ``` ```bash conda env create -f environment.yml conda activate brazil-pix-replication ``` ## Path 1: Frozen Processed Data Run this one-liner from the package root: ```bash python3 code/run_cleaned.py --output outputs_from_cleaned && python3 code/check_outputs.py --outputs outputs_from_cleaned ``` This path uses `data/cleaned/analysis_data.parquet` and the other frozen processed files in `data/cleaned/`. It reproduces all final-paper JFSR tables and figures in: - `outputs_from_cleaned/jfsr_tables/` - `outputs_from_cleaned/jfsr_figures/` - `outputs_from_cleaned/jfsr_docs/` The online appendix is reproduced as `outputs_from_cleaned/jfsr_docs/online_appendix.docx`; its underlying appendix tables and figures are stored in `jfsr_tables/` and `jfsr_figures/` with `online_appendix_` prefixes. Expected runtime on a modern laptop is roughly 30-60 minutes. ## Path 2: Uploaded Raw Public Data Upload every file listed in `docs/raw_files_to_upload.csv` to the listed `required_upload_path`, preserving filenames and folders. Then run this one-liner from the package root: ```bash python3 code/run_raw.py --output outputs_from_raw && python3 code/check_outputs.py --outputs outputs_from_raw ``` The raw upload list contains 204 public files totaling about 835 MB. The raw path rebuilds the municipality-month analysis panel and then reproduces the same final-paper and online-appendix outputs. ## Output Checks `code/check_outputs.py` compares generated JFSR tables and online-appendix tables against `reference_outputs/jfsr_tables/` and verifies required figure files, the online appendix DOCX, the JFSR table/figure crosswalk, the in-text-number file, and the output checksum file.