Multilateral Tariff Diversion in United States Machinery and Electronics Imports: A Continuous Difference-in-Differences Approach
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Evaluating 36 consecutive months of bilateral U.S. import records from the UN Comtrade Database across machinery and electronics (Harmonized System Chapters 84 and 85; N = 25,544 positive trade observations in the baseline estimation panel across 133 four-digit HS headings), this paper investigates global supply chain reallocation following 2025 United States reciprocal border duties. Because official customs declarations record strictly cross-border merchandise imports rather than domestic factory shipments, our empirical design evaluates sourcing reallocation among foreign suppliers and cannot rule out domestic reshoring. Under Chapter 99 provisions (Annex II secondary subheading 9903.01.32), non-China shipments under exempt technology categories entered duty-free under longstanding Information Technology Agreement (ITA-1) rules, while Chinese shipments faced sharp discretionary tariff escalation (peaking at 145% from April 10 to May 13 before the May 14 Geneva truce under Executive Order 14298 lowered China’s reciprocal rate to 10%, alongside a 20% IEEPA emergency duty on exempt hardware). Implementing a continuous difference-in-differences design using an exogenous pre-panel baseline (ChinaShare_s^2022), we estimate a multilateral diversion semi-elasticity relative to China of β̂_ExChina = 1.4720 (SE = 0.2886, p < 0.0001; two-way clustered p = 0.0036 under t(5)). Parametric trend-break splines and dynamic event studies establish pre-treatment baseline stability followed by a discrete post-announcement enforcement step jump (δ̂ = 1.3673, p < 0.0001). Sourcing reallocation extends broadly across capital goods outside computing and telecommunications apparatus (β̂_NonTech = 1.2095, p = 0.0001). An exact additive quantity–price decomposition establishes that 91.3% of relative dollar reallocation reflects physical shipping tonnage rather than border price escalation, while landed unit values exhibit a statistically imprecise positive response (β̂_P = +0.1181, SE = 0.0842, p = 0.1631; 95% CI [−0.048, +0.285]), lacking statistical power to rule out substantial exporter markups or distinguish volume expansion from invoice relabeling.
Fields of Study & Subjects
Global trade and economics
COVID-19
Geopolitics
Technology
Migration
Global trade
sustainability
and social impact
Publication details
- Journal / source
- Munich Personal RePEc Archive (Ludwig Maximilian University of Munich)
- Publication date
- 27 March 2026
- Language
- EN