Replication Package for “Trade Policy, Industrial Subsidies, and the Structural Trap: Policy Shock Evidence from India’s Product Space”
Description
This replication package contains all data and R code required to reproduce the empirical results in Vhankade and Gaur (2026), “Structural Transformation, Institutional Constraints, and Export Upgrading: Product Space Evidence from India,” submitted to the Journal of Development Economics. The paper constructs a monthly Structural Change Index (SCI) for India using commodity-level export data from the Directorate General of Commercial Intelligence and Statistics (DGCI&S) and Product Complexity Index values computed from the Harvard Growth Lab Atlas. It estimates the effects of India’s RCEP exit (November 2019) on import penetration using a triple-difference design, and the effects of Production Linked Incentive schemes on import substitution using a difference-in-differences design. The package includes R scripts for all analyses, computed intermediate datasets, and a README with step-by-step replication instructions. The DGCI&S trade micro data require a separate access application to the Government of India; the package provides the derived datasets and all publicly available inputs.
Files
Steps to reproduce
C1. Software Requirements-● Install all packages by running: install.packages(c("dplyr","readr","lubridate","tidyr","stringr", "fixest","strucchange","ggplot2","economiccomplexity", "zoo","did")) C2. Directory Setup-Create the following directory structure on your machine: ~/replication/ ~/replication/scripts/ ~/replication/data/raw/ ~/replication/data/derived/ ~/replication/external_data/pci_raw/ ~/replication/external_data/figures_v2/ C3. Data Acquisition-https://dgciskol.gov.in/. Upon receiving access, save the workspace file as: ~/replication/India_Trade_Workspace.RData Step 2 — Harvard Atlas HS12 6-digit data (public):-https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/XTAQMC Step 3 — Verify all raw files are in place: C4. Run the Replication Pipeline- Option A — Full automated pipeline (requires DGCI&S access): Option B — Partial replication without DGCI&S access: Option C — Direct verification of key outputs: C5. Expected Outputs and Verification ● sci_monthly.rds — 100 rows, columns: ym, SCI, n_products, total_value ● sci_annual.rds — 9 rows (years 2015-2025 excluding 2022-2023) ● pci_hs12_6digit.rds — 5,204 rows, columns: hs6_code, pci, hs2_code, pci_rank, pci_pctile ● master_panel.rds — 20,056,969 rows ● bai_perron_results.rds — breakpoint test results with supF_break_date = 2017-06-01 ● rcep_triple_diff_results.rds — fixest model objects m1 through m4 ● pli_did_results.rds — fixest model objects m_pli1 through m_pli4, m_pli_sector ● policy_simulation_results.rds — scenario_A_deficit = 120.55, scenario_B_deficit = 192.16 Key regression coefficients to verify against Table 3 and Table 4 in the paper: # RCEP DD (Table 3, col 1): should be 0.0645 (SE 0.0271) # RCEP DDD triple (Table 3, col 2): should be -0.1008 (SE 0.0369) # PLI pooled (Table 4, col 1): should be -0.0080 (SE 0.0442) # PLI x CD_high (Table 4, col 2): should be -0.1128 (SE 0.0513) C6. Correspondence Questions about the replication package should be directed to Dr. Prakash G. Vhankade at prakash.vhankade@gipe.ac.in. Questions about access to DGCI&S trade microdata should be directed to the Directorate General of Commercial Intelligence and Statistics, Government of India (https://dgciskol.gov.in/). SECTION D: FILE-LEVEL DESCRIPTIONS D1. Script Files 01b_compute_pci.R: Computes Product Complexity Index (PCI) for 5,204 HS6 products using the Method of Reflections (Hidalgo and Hausmann, 2009) applied to Harvard Atlas HS12 6-digit bilateral trade data for base year 2022 (N = 229 countries). Output: pci_hs12_6digit.rds and pci_hs12_6digit.csv. 02_construct_SCI.R: Constructs the monthly Structural Change Index (SCI) by merging PCI values with DGCI&S export observations and computing value-weighted mean PCI per month. Also constructs complexity distance (CD), CD_high binary indicator, and CD_quartile. Outputs: sci_monthly.rds, sci_annual.rds, complexity_distance.rds. 03_build_institutional_variables.R: 04_bai_perron_SCI.R
Institutions
- Gokhale Institute of Politics and EconomicsMaharashtra, Pune