Replication Data for: Procurement Certainty, Coordination Amplification, and Acreage Lock-in: Graduated Intensity Identification and Panel Quantile Evidence from Maharashtra, India

This replication package contains the data and R code required to reproduce all tables and results reported in the manuscript "Procurement Certainty, Coordination Amplification, and Acreage Lock-in: Graduated Intensity Identification and Panel Quantile Evidence from Maharashtra, India" (submitted to Journal of Quantitative Economics). The package contains five files: 1. RODE_Analysis_Panel_238obs.csv Analysis dataset. Balanced panel of 238 district-year observations covering 34 agricultural districts of Maharashtra across seven Kharif seasons (2016-17 to 2022-23). RODE_01_Codebook.R. 2. RODE_01_Codebook.R Documents all variables including identifiers, acreage share dependent variables, the continuous cooperative mill density variable (log_mills), rainfall SPI-3 instrument, structural controls, PMFBY state-level control, and AgriStack DiD indicators. 3. RODE_02_Build_Dataset.R Data construction script. Builds RODE_Analysis_Panel_238obs.csv from six primary Indian government administrative data sources: (i) Directorate of Economics and Statistics, MoAFW (data.desagri.gov.in); (ii) India Meteorological Department district daily rainfall series; (iii) Maharashtra Cooperative Sugar Factories Federation official factory registry (mahasugarfed.org); (iv) PMFBY Administrative Dashboard (pmfby.gov.in); (v) Ministry of Agriculture and Farmers Welfare Digital Agriculture Mission rollout records (agristack.gov.in); (vi) Companion district panel constructed from Maharashtra Season and Crop Reports and ICRISAT district-level database. 4. RODE_03_Estimation.R Estimation script. Reproduces all regression results reported in Tables 3 through 6 of the manuscript: panel fixed effects with continuous mill density (dose-response), coordination amplification interaction specification, panel quantile regression using the Canay (2011) two-step estimator, two-way fixed effects DiD with pre-trend tests, Generalized Synthetic Control estimation (Xu, 2017) via the gsynth package, and robustness specifications. Requires R packages: plm, lmtest, sandwich, quantreg, gsynth, tidyverse. 5. RODE_04_README.txt Documentation. Describes data sources, sample construction, variable definitions, missing value handling, AgriStack cohort assignments, estimation steps, and known limitations. Required reading before running any script. All underlying crop acreage and production data are publicly accessible through Indian government portals as documented in RODE_02_Build_Dataset.R and RODE_04_README.txt. Software: R version 4.4.3 (aarch64-apple-darwin20). Key package versions: plm 2.6-3, quantreg 5.97, gsynth 1.2.1. Prakash Gurumurti Vhankade, Associate Professor, Gokhale Institute of Politics and Economics, Pune, [e-mail]