Measuring the Effect of Tariffs on Critical Minerals Trade in a Multidimensional Panel
Abstract Estimated tariff elasticities for critical minerals change several-fold as fixed effects are added and this instability is usually treated as specification sensitivity. This paper traces it to non-identification. A tariff schedule published for an importer, a product and a year takes one value for every supplier, so in a structural gravity system with importer-product-year effects the rate cancels from the gravity equation and its coefficient describes a parameter absent from the model. The remedy is a rate that differs across suppliers through preferential access. A Poisson pseudo-maximum likelihood gravity model with exporter-mineral-year, importer-mineral-year and pair-mineral effects on a multi-dimensional panel of bilateral mineral trade gives a negative tariff elasticity, significant under pair-clustered inference, much smaller than additive-effects estimates and little changed by a bias correction. Log-linear, nonparametric and interactive-effects estimators find no significant pooled response. Trade in cobalt, copper and graphite falls with the tariff, trade in lithium rises and the rare earths show no response. The residual variance share shows that few flows carry a supplier-specific rate, so the estimate is an association identified from few cells and not a general elasticity. It is not significant within short subperiods and the interval includes zero under a pair bootstrap. JEL Classification: C23, F13, F14
Authors
- Muhammad Afnan Arif (ORCID: https://orcid.org/0009-0008-9717-5275)
Institutions
- University of Malaya (MY)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-13
- DOI
- https://doi.org/10.21203/rs.3.rs-11503075/v1
- Primary Topic
- Global trade and economics
- Type
- preprint