Replication data: agricultural eco-efficiency in Latin America and the Caribbean, 1995-2020 (FAOSTAT and World Bank WGI, calibrated fuzzy sets for fsQCA)
Description
Country-level dataset supporting the study "Configurational paths to agricultural eco-efficiency in Latin America and the Caribbean (1995-2020): a two-period fuzzy-set qualitative comparative analysis". It covers 27 Latin American and Caribbean countries over 1995-2020 (702 country-year records) and two 13-year periods, 1995-2007 and 2008-2020 (54 country-period observations). Raw inputs come from two public sources: FAOSTAT (FAO) for fertilizer nutrients, feed use, agricultural area, agricultural labor, net capital stock, livestock manure (kg of nitrogen), gross production value and agricultural greenhouse-gas emissions; and the World Bank Worldwide Governance Indicators, Government Effectiveness: Estimate. The workbook builds, with live formulas, the eco-efficiency outcome (gross production value per kt CO2-equivalent), three production-regime conditions (fertilizer per hectare, feed per unit of manure nitrogen, capital per worker) and state capacity (government effectiveness), then calibrates them into fuzzy sets. It also holds the common calibration anchors (5th, 50th and 95th percentiles of the 54 pooled observations) and two matrices ready to import into fsQCA software, one per period (also provided as CSV). Note: the World Bank WDI release used (updated 29 October 2025) reports Government Effectiveness on an expanded scale (about -32 to +24) instead of the canonical -2.5 to +2.5; it is the same indicator rescaled (Pearson 0.996 and Spearman 0.995 against the original WGI series over 3,624 overlapping observations), so percentile-based calibration is unaffected. Annual Government Effectiveness is missing for 1995 and biennial until 2002; gaps were left empty and not imputed. The README sheet and the replication protocol sheet are written in Spanish; variable names and units are self-explanatory.
Files
Steps to reproduce
1. Download from FAOSTAT, for the 27 countries and years 1995-2020, eight variables: fertilizer by nutrient (t), feed (t), agricultural area (thousand ha), labor (thousand persons), net capital stock (million constant 2015 USD), livestock manure (kg N), gross production value (thousand constant 2014-2016 international $) and agricultural emissions (kt CO2e). Download Government Effectiveness: Estimate from the World Bank WGI (https://api.worldbank.org/v2/en/indicator/GE.EST?downloadformat=excel); leave missing years empty. Sheet: Brutos_anuales. 2. Compute annual ratios: eco-efficiency = gross production value / emissions; fertilizer intensity = fertilizer / agricultural area; confinement = feed / manure N x 1000; capital intensity = net capital stock / labor; institutional capacity = government effectiveness as is. Sheet: Calculadas_anuales. 3. Average each ORIGINAL variable within each period (1995-2007, 2008-2020), not the ratios. Sheet: Brutos_periodo (27 countries x 2 periods = 54 rows). 4. Recompute the five ratios on the period means. Sheet: Calculadas_periodo. 5. Take natural logarithms of eco-efficiency, fertilizer intensity, confinement and capital intensity; government effectiveness is not transformed. 6. Set common anchors for both periods as percentiles 5 (full exclusion), 50 (crossover) and 95 (full membership) of the 54 pooled observations. Sheet: Anclajes. 7. Calibrate with the logistic function: z = (x - crossover) x ln(19) / (full - crossover) if x >= crossover, or (x - crossover) x ln(19) / (crossover - exclusion) otherwise; membership = 1 / (1 + exp(-z)). Sheet: Fuzzy_periodo. Import the two matrices (sheets fsQCA_P1_1995_2007 and fsQCA_P2_2008_2020, or the CSV files) into fsQCA software for the truth-table analysis. Checks: no membership outside [0,1] or empty, none exactly 0.5.
Institutions
- Universidad de SevillaAndalusia, Seville