Value Investing in Brazil: Long-Term Returns Relative to the Market Benchmark
Description
This dataset provides a complete replication package for the empirical results reported in the associated study on a rules-based, Graham-style equity screening strategy in the Brazilian stock market. It includes self-contained Python scripts and companion Excel workbooks that reproduce the main statistical tests, robustness analyses, and benchmark comparisons. The package contains: (i) annual and monthly inference tests (including heteroskedasticity/autocorrelation-consistent inference and resampling-based confidence intervals), (ii) risk-adjusted performance analysis via time-series factor regressions using publicly available NEFIN factor data, (iii) robustness tables based on sensitivity checks and quintile-style analyses, and (iv) a transparent transaction-cost sensitivity analysis (bps per trade) reporting the impact on cumulative performance metrics. All calculations are provided in verifiable Excel workbooks and standalone Python scripts to facilitate auditability and replication. The scripts are designed to run offline (no external downloads required), and outputs are generated in spreadsheet format for straightforward inspection. Data sources: The factor series used for risk-adjustment are based on NEFIN (Núcleo de Estudos em Finanças, FEA-USP). Benchmark rate series (e.g., Selic) are based on publicly available official sources.
Files
Steps to reproduce
Steps to reproduce 1) Requirements Python 3.10+ (recommended) Packages: numpy, pandas, scipy, statsmodels, openpyxl Install dependencies (example): pip install numpy pandas scipy statsmodels openpyxl 2) Run the full replication (recommended) Run: python annual_tests_suite.py This produces/updates the main statistical-test outputs in: annual_tests_workbook.xlsx 3) Risk-adjusted performance (factor regressions) Run: python nefin_factor_regressions.py This produces/updates: nefin_factor_regressions.xlsx 4) Robustness tables (sensitivity and quintiles) Run: python robustness_tables_builder.py This produces/updates: robustness_quintiles.xlsx 5) Transaction-cost sensitivity analysis Run: python trading_cost_sensitivity.py This produces/updates the cost sensitivity outputs (tables) and can be cross-checked against: _strategy_vs_ibov_selic.xlsx 6) Optional: inference workbook generation If you want to regenerate the consolidated inference workbook (templates + tests): python inference_tests_builder.py This produces/updates: inference_tests_workbook.xlsx 7) Verification After running the scripts, open the Excel workbooks and compare the generated tables with the reported results in the associated manuscript. All workbooks are provided in a human-auditable format.
Institutions
- Faculdade FIPECAFI