Investment as a Substitute for Governance: A Configurational Analysis of Growth Across BRICS and Partner Economies — Replication Package
Description
This dataset provides the complete replication package for “Investment as a Substitute for Governance: A Configurational Analysis of Growth Across BRICS and Partner Economies” (manuscript targeted at South Asian Journal of Business Studies, Emerald). The study applies fsQCA and NCA to a BRICS-plus sample expanding across three documented stages (N=5 founding BRICS, N=11 BRICS Plus, N=15 with an Asian partner-tier addition; N=20 reported as a robustness extension), built from World Bank Worldwide Governance Indicators and World Development Indicators retrieved live 20–24 August 2026. It tests whether a low-governance/high-investment configuration outperforms the conventional high-governance/high-investment path in sufficiency consistency, and whether investment carries the only significant NCA necessity effect. The deposit contains: (1) four raw World Bank panel CSVs plus intermediate country-mean CSVs; (2) 15 analysis scripts (R and Python) — canonical R scripts (QCA/NCA packages) for every cited number, Python scripts retained as independent design-stage cross-checks, clearly marked as non-citable; (3) SESSION_INFO.txt recording exact R/Python/package versions; (4) a README with a full script-to-manuscript-location mapping table covering the canonical truth table, NCA necessity/bottleneck analysis, PRI-cut robustness, the N=15/N=20 expansion, the governance-proxy robustness check (Control of Corruption vs. Government Effectiveness vs. Regulatory Quality), the crisp-set (csQCA) robustness check, and the human-capital rival-condition check.
Files
Steps to reproduce
Install R 4.3.3 or later with the QCA (v3.25) and NCA (v5.0.2) packages: R -e 'install.packages(c("QCA","NCA"))'. Install Python 3.11 or later with pandas, numpy, scipy, and requests: pip install pandas numpy scipy requests. Keep all files in a single flat directory — every script reads its input CSVs by bare filename from its own directory, and two scripts (new5_data.py, human_capital_data.py) are imported as modules by others, so do not reorganize into subfolders. The four raw World Bank panel CSVs (wb_governance_panel.csv, wb_growth_panel.csv, brics_plus_new_members_panel.csv, brics_partner_asia_panel.csv) are static inputs already included; no script generates them. Run generate_new5_country_means.py and ge_rq_data.py first if regenerating their output CSVs from scratch (pre-computed copies are already included, so this step is optional). Run the canonical R scripts to reproduce every cited number: paper_b_canonical_QCA.R (Table 2a/2b truth table and Boolean minimization), paper_b_canonical_NCA.R (Table 3 necessity effects), bottleneck_table.R (Table 3a bottleneck table), human_capital_canonical.R (§4.3 human-capital check), alt_proxy_canonical.R (governance-proxy robustness), crisp_set_robustness.R (crisp-set robustness), and n20_canonical.R (N=20 extension). The Python scripts (paper_b_truthtable.py, paper_b_nca.py, and others) are independent design-stage cross-checks, not the source of any cited number — run them only to see the hand-written verification logic, not to reproduce manuscript figures. Compare your output against SESSION_INFO.txt’s recorded package versions if any number doesn’t match to the last decimal.
Institutions
- University of LucknowUttar Pradesh, Lucknow