Mirror statistics compare what one country reports exporting with what its partner reports importing. The version below is a variant: rather than comparing two reports of the same flow, it puts two different flows side by side — what Germany and the United States report shipping into a transit economy, and what that transit economy reports shipping onward to Russia.
Both series come from UN Comtrade, both are annual, and both cover Harmonised System chapters 84 and 85 — machinery and mechanical appliances, and electrical machinery and equipment. Those two chapters contain most of the items that appear on published high-priority commodity code lists.
Armenia
Germany and the United States into Armenia Armenia onward to Russia
The same figures as a table
| Year | DE + US into Armenia | Armenia to Russia | Onward as share of inbound |
|---|---|---|---|
| 2019 | $60m | $14m | 23% |
| 2020 | $176m | $44m | 25% |
| 2021 | $128m | $55m | 43% |
| 2022 | $321m | $1.3bn | 400% |
| 2023 | $296m | $1.2bn | 419% |
| 2024 | $255m | $1.2bn | 468% |
Armenia reported exporting about fifty-five million dollars of these chapters to Russia in 2021. In 2022 it reported one and a quarter billion — a factor of twenty-three, in one year, in an economy of about three million people.
Kyrgyzstan
Germany and the United States into Kyrgyzstan Kyrgyzstan onward to Russia
The same figures as a table
| Year | DE + US into Kyrgyzstan | Kyrgyzstan to Russia | Onward as share of inbound |
|---|---|---|---|
| 2019 | $20m | $5.7m | 29% |
| 2020 | $28m | $7.9m | 28% |
| 2021 | $34m | $7.9m | 23% |
| 2022 | $137m | $85m | 62% |
| 2023 | $383m | $235m | 61% |
| 2024 | $422m | $208m | 49% |
Kyrgyzstan’s onward exports in these chapters ran at six to eight million dollars a year through 2021. By 2023 the figure was two hundred and thirty-five million, roughly thirty times the pre-2022 level, and inbound German and US shipments had risen more than tenfold over the same period.
Kazakhstan
Germany and the United States into Kazakhstan Kazakhstan onward to Russia
The same figures as a table
| Year | DE + US into Kazakhstan | Kazakhstan to Russia | Onward as share of inbound |
|---|---|---|---|
| 2019 | $866m | $275m | 32% |
| 2020 | $1.5bn | $204m | 14% |
| 2021 | $1.4bn | $251m | 18% |
| 2022 | $2.2bn | $1.7bn | 74% |
| 2023 | $3.2bn | $1.8bn | 56% |
| 2024 | $2.9bn | $1.6bn | 56% |
Kazakhstan is the largest of the three by some distance and shows the same shape at a different scale: onward exports of about two hundred and fifty million dollars a year before 2022, rising to between one and a half and one and three quarter billion afterwards.
What this does and does not establish
It establishes that the flows changed, sharply, in the year direct routes closed, in commodity chapters that export control authorities had specifically identified. That is a real finding, it comes from the countries’ own customs reporting, and anyone can reproduce it.
It does not establish that any particular shipment was unlawful, that any particular company did anything, or that all of the increase is re-routing. Several things can produce part of this pattern legitimately:
- Genuine demand growth. Companies relocated staff and operations to these countries in 2022, and those operations import machinery.
- Customs union effects. Armenia, Kazakhstan and Kyrgyzstan are all in the Eurasian Economic Union with Russia, and intra-union trade reporting has its own conventions.
- Re-export as an ordinary business. Entrepôt trade is a legitimate economic activity and these economies do it.
- Statistical revision. Comtrade figures are revised, and the most recent year is the least settled.
What the size of the change does is shift the question. A doubling invites several explanations. A twenty-three-fold increase in one year, concentrated in the chapters an export control authority named, narrows the field considerably — and it is why these three countries appear in every serious analysis of this subject, including analysis published by their own governments.
The method
The query, the exact endpoint, the filters applied and the reasons for each are set out in full in the methodology note. The underlying data is refetched monthly by a script in this repository and the last fetch date is recorded on the sources page.