Global Arms Procurement Timing and Access Dataset (GAPTAD), 2000–2019
Description
The Global Arms Procurement Timing and Access Dataset (GAPTAD), 2000–2019 is a transaction-level dataset developed for research on the timing, completion, congestion and political conditions of international arms procurement. It contains 7,502 records covering 151 recipient countries, 87 supplier countries and 11 major weapon categories. When recipient and supplier roles are combined, the dataset includes 155 unique countries. The unit of observation is a recipient–supplier–weapon transaction or order. Each record includes the recipient and supplier, weapon description and category, order year, ordered and delivered quantities, first and last observed delivery years, delivery delays, completion ratio and pending or incomplete status. This structure allows researchers to distinguish first delivery from full order completion and examine whether large orders begin delivery quickly but remain incomplete for longer periods. The dataset is built around major conventional arms-transfer records originating from the SIPRI Arms Transfers Database and is harmonised with country-year and country-pair indicators from public international sources. It includes measures of supplier-category backlog, same-year competing orders, bilateral political distance based on UN General Assembly voting, political alignment, supplier dependence, previous trade relationships, regional conflict exposure, military expenditure and real GDP. SIPRI Trend-Indicator Values are used only as volume indicators and must not be interpreted as prices, contractual values or verified payments. The dataset does not identify whether a system was fully paid for, the timing of payments or the definitive political reason for suspension or non-delivery. Indicators for embargoes, export-licence restrictions, political interruptions and cancellations should therefore be treated as reproducible signals rather than conclusive causal classifications. Regional threat measures combine conflict fatalities outside the recipient country with geographical distance, allowing security exposure to be represented even when the country is not directly at war. The final analytical file contains 51 variables, each abbreviated to no more than four characters. A codebook explains definitions, sources, formulas and coding rules. No substantive economic, political, conflict or transaction value was statistically imputed. Where delivery dates were not observed, documented sentinel codes were used instead of invented dates. Users should consult the codebook before recoding these values. GAPTAD supports descriptive analysis, fixed-effects estimation, binary-response models, survival analysis, competing-risk frameworks, network analysis and weapon-category heterogeneity studies. It is intended to improve transparency and reproducibility in research on defence procurement, delivery delays, order completion, supplier dependence, production bottlenecks and the political economy of international arms transfers.