Global Procurement Research Dataset (GPRD): Cross-Country Procurement Data and Causal Analysis Toolkit

The Global Procurement Research Dataset (GPRD) harmonizes 2.3+ million public procurement contract records from Ukraine (ProZorro), Colombia (SECOP), and the UK (Contracts Finder) using the Open Contracting Data Standard (OCDS) v1.1. This dataset supports causal research on how procurement rules affect competition, prices, quality, and innovation. Key features: **Data Sources:** - Ukraine ProZorro: 1.2M contracts (2016-2024) - Colombia SECOP II: 680K contracts (2015-2024) - UK Contracts Finder: 420K contracts (2016-2024) **Identification Strategies:** 1. Regression Discontinuity Designs (RDD) at monetary thresholds where procurement rules change discretely 2. Difference-in-Differences (DiD) analysis of e-procurement reform adoption 3. Event studies with Callaway-Sant'Anna staggered adoption estimators **Novel Features:** - "Mechanism Index": NLP-based measure of contract restrictiveness, complexity, and innovation orientation from tender text - PatentsView linkage for supplier innovation capacity - Cross-country harmonized CPV/UNSPSC sector classifications - Threshold indicators for EU and national procurement regulations **Contents:** 1. src/ - Complete Python pipeline for data acquisition, harmonization, causal analysis 2. Data/ - Raw and processed procurement records from all three countries 3. DATA_CODEBOOK.md - Documentation for 80+ GPRD variables 4. Research_paper/paper.tex - Full manuscript (Nature Human Behaviour format) 5. requirements.txt - Python dependencies **Key Findings:** - 15.2pp increase in competitive bidding above publication thresholds - 8.3% reduction in contract prices - 23% lower text restrictiveness in above-threshold contracts - Suppliers of above-threshold contracts hold 2.4× more patents Associated manuscript targets Nature Human Behaviour / Nature Sustainability.