Replication Data for "Fertilizer Policy Reforms in the Midst of Crisis: Evidence from Rwanda"

This dataset is provided to support the replication of the results presented in the paper <i> "The Impact of Fertilizer Subsidies on Agricultural Productivity in Rwanda" </i> (Spielman, David J.; Mugabo, Serge; Rosenbach, Gracie et al., forthcoming). It includes both the raw data and the processed variables used in the analysis.</p>

The data for this study is sourced from the National Institute of Statistics of Rwanda's (NISR) Seasonal Agriculture Survey (SAS), conducted between 2017 and 2020. The dataset includes a representative sample of small-scale and large-scale farms across Rwanda. It contains detailed information on land use, crop production, fertilizer usage, crop yields, and agricultural practices across the country’s three agricultural seasons: Season A (September to February), Season B (March to June), and Season C (July to September).</p>

The SAS dataset includes data from 76,182 sampled plots across Rwanda, capturing key agricultural variables such as fertilizer application, crop types, land sizes, and yields. This dataset is further enriched with additional data from the National Agriculture Export Board (NAEB), which provides information on tea and coffee production through routine monitoring and a five-year coffee census. Input subsidies, pricing, and distribution data are sourced from the Ministry of Agriculture and Animal Resources (MINAGRI) and the Rwanda Agriculture and Animal Resources Board (RAB).</p>

Accompanying this dataset are the do-files, which contain the Stata code necessary to replicate the analysis presented in the paper. These scripts include all steps for data cleaning, analysis, and result generation. Detailed, step-by-step instructions for replicating the analysis can be found in the included README file.</p>

While the dataset is based on the original SAS data, it has been cleaned and transformed for analysis. Missing values have been handled using mean imputation for continuous variables and mode imputation for categorical variables. Outliers in crop yield data have also been adjusted based on established thresholds.