An Eightfold Aggregation Trap in Mirror Trade Evidence on Third-Country Rerouting of Controlled Goods
Commentary and Critique · 2026
DOI: 10.62432/appe.cc.2026.9.34 ↗Kurz gefasst
Fachartikel in der Zeitschrift Commentary and Critique, 2026. Verfasst von Farhod Rakhimov und 1 weiteren Autor:innen. Die Arbeit nennt 7 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Third-country rerouting of export-controlled goods is a documented procurement pattern, and the temptation is to evidence it from mirror trade statistics — the gap between what an exporter reports sending and what the importing partner reports receiving. This article tests whether that gap can separate rerouting of controlled goods from ordinary statistical attribution differences, using the United Nations Comtrade public preview interface, the joint Common High Priority Items List as the operable definition of "controlled", and a same-chapter off-list contrast. It reports two results. The first is a measurement result that any user of the interface needs: the endpoint's response for a single reporter–partner–commodity–year–flow cell is **not an additive cross-tabulation**. Across 160 non-empty cells the naive sum of a cell's rows divided by that cell's maximum is approximately **8.00 within the recorded precision in 133 cells (83.1%)**; the total-bearing row is the one with `customsCode=C00` **and** `motCode=0` **and** `partner2Code=0` in **160 of 160 cells**, yet no field in the response marks it: the descriptor fields are null in all 9,790 returned records. The factor of eight is approximate at raw floating-point precision. It decomposes into a 2.00 partner-2 multiplier and a 4.00 mode-by-customs multiplier across the affected cells; the worked example shows the latter as two 2.00 components. The second result concerns the controlled-versus-off-list comparison. On the eight mirror relations this design can build (2023, Germany↔Kazakhstan, four listed and four off-list commodities), the median absolute log value gap is **0.620** — a median percentage gap of **+85.9%** — and the controlled group does **not** separate from the off-list group: exact two-sided Mann-Whitney U test p = **0.343**, permutation p = **0.314**, n = 4 versus 4. With four observations per group, this comparison detects no difference and has little power to distinguish plausible effects. The weight dimension, which carries no valuation basis, shows a median gap of **+93.5%**, but in six of the eight relations one leg's net weight is publisher-imputed, so that comparison is reported under a flag and is not used as the decisive argument. On the data this design can reach, mirror gaps do not isolate rerouting; testing the effect requires a larger controlled comparison and evidence of physical movement beyond mirror records.
Zitieren
APA
Rakhimov, F., & Qodirova, N. (2026). An Eightfold Aggregation Trap in Mirror Trade Evidence on Third-Country Rerouting of Controlled Goods. Commentary and Critique. https://doi.org/10.62432/appe.cc.2026.9.34
Vancouver
Rakhimov F, Qodirova N. An Eightfold Aggregation Trap in Mirror Trade Evidence on Third-Country Rerouting of Controlled Goods. Commentary and Critique. 2026. doi:10.62432/appe.cc.2026.9.34
BibTeX
@article{Rakhimov2026,
author = {Rakhimov, Farhod and Qodirova, Nargiza},
title = {An Eightfold Aggregation Trap in Mirror Trade Evidence on Third-Country Rerouting of Controlled Goods},
journal = {Commentary and Critique},
year = {2026},
doi = {10.62432/appe.cc.2026.9.34}
}
EUVIMED