Replication data and source files for Economic, Social, Environmental, and Institutional Factors Shaping Onshore Wind LCOE in Europe: Evidence from Spatial Panel Models
Description
This replication package accompanies the article “Economic, Social, Environmental, and Institutional Factors Shaping Onshore Wind LCOE in Europe: Evidence from Spatial Panel Models.” It tests the hypothesis that how economic, social, environmental, and institutional macro-level factors shape onshore wind LCOE in Europe via a Spatial Durbin Model with individual fixed effects. The dataset is a balanced annual panel of 14 OECD European countries (2003-2022). In the associated paper, spatial spillovers dominate domestic channels. Neighboring-country GDP and GHG are associated with lower domestic LCOE (via spillovers), whereas HDI and GRI are positively associated with LCOE both domestically and through spillovers—implying policy gains from regional coordination. LCOE for onshore wind (constant 2023 USD/kWh) is from IRENA; GDP per capita is from the World Bank’s WDI; HDI and WGI are obtained via the QoG Standard Dataset (which compiles UNDP’s HDI and the World Bank’s WGI); GHG emissions per capita are from the 2024 Environmental Performance Index (Yale University). All monetary series are converted to constant 2023 USD. Also, variables are provided in natural logs (e.g., Ln.LCOE, Ln.GDP, Ln.HDI, Ln.GHG, Ln.GRI). To obtain a balanced panel, monotonic cubic interpolation is used to fill in the missing values between two observed years, which accounts for 47 missing observations. Moreover, a forward filling approach is used to extrapolate limited future gaps based on the last available value from earlier years, which fills in 7 additional missing observations. In addition to the panel data, two types of spatial information are employed to support the econometric analysis. First, two polygon-based shapefiles are obtained from the World Bank, one in low resolution and one in high resolution. The high-resolution shapefile is used for all econometric estimations, while the low-resolution version is reserved for thematic mapping only. Both shapefiles are projected to a coordinate reference system optimized for Europe (EPSG:3035). Additionally, capital city centroids are used as spatial reference points when constructing spatial linkages across countries. These centroids are collected from the countryref dataset in the CoordinateCleaner package in R. Please refer to the "README.md" file for a quick start and an end-to-end workflow overview. The code is self-contained and self-explanatory. Running the "_main.R" file reproduces all tables and figures reported in the paper. No synthetic data is generated, and all paths are relative to the current directory. For methodological details and interpretation, refer to the manuscript. All compiled data and code in this package are shared under the license stated on this record. Raw third-party data remain available from the cited providers; links and exact retrieval/processing instructions are explained in the "README.md" file.
Files
Steps to reproduce
The dataset is assembled from multiple publicly available international sources. Specifically, raw data is retrieved from the following repositories: International Renewable Energy Agency (IRENA) for onshore wind LCOE; World Bank World Development Indicators (WDI) for GDP per capita; Quality of Government Institute (QoG) Standard Dataset for the Human Development Index (HDI) and the World Governance Indicators (WGI), which QoG compiles from UNDP and the World Bank respectively; the 2024 Environmental Performance Index (EPI) for greenhouse gas emissions per capita from Yale University. Geospatial files include country polygon shapefiles obtained from the World Bank (in both low- and high-resolution formats), as well as capital-city centroids sourced from the countryref dataset in the R package CoordinateCleaner. Workflow: 1. Raw data collection: Each dataset is downloaded in its original format and stored in source-specific subfolders. For transparency, every subfolder includes a “—Data Retrieved.txt” file (containing the retrieval date and source link), a “—Notes.docx” file (with collection notes), and the raw data file. 2. Data ETL: The script "load_data.R" processes all raw data sources. It performs cleaning, transformations, and merging before exporting processed datasets into the "Output/_Data" directory. This script is automatically called by `_main.R`. Users may also run "load_data.R" independently to reproduce the processed dataset from raw sources. 3. Panel dataset construction: A balanced panel dataset is created for 14 European countries. Datasets are merged into a single file (“combined.csv” and “combined.RData”) stored under the "Output/_Data/_Combined" directory. 5. Reproducibility: All steps are scripted in R and automated via the main script. The R package environment is pinned using the "renv" folder and “renv.lock” file. A Stata folder is also provided with the necessary files for performing the final phase of the analysis in Stata. Software: R (version 2025.05.0+496) and Stata (version 17 or later). All analyses can be fully reproduced by restoring the R environment (renv::restore) and sourcing the "_main.R" script.
Institutions
- Eskisehir Osmangazi Universitesi