Re-export routes through Central Asia and the Caucasus: what open sources can and cannot show

Since 2022, trade between Russia and its neighbours in Central Asia and the South Caucasus has drawn sustained attention from regulators, compliance teams and journalists. The concern is that goods restricted by Western sanctions and export controls may reach Russia through third countries. Open sources can help. They can also mislead, and the distance between a statistical pattern and evidence about a specific company is wide.

What mirror statistics show

The core method is the comparison of mirror statistics. Most countries report their imports and exports by partner and product, and these figures are collected in databases such as UN Comtrade. In principle, what one country records as exports to a partner should resemble what the partner records as imports from it. Comparing the two sides of the same flow can reveal gaps, sudden changes and patterns that one country's data alone would hide.

Mirror data has become more important because Russia has published much less detailed trade data of its own since 2022. Researchers therefore often reconstruct its trade from what partners report. A typical exercise compares a neighbouring country's imports of a product from Europe or Asia with its exports of the same product to Russia, over a period that spans 2022. Some countries report late, incompletely or not at all, and that should be recorded as a limitation.

Membership of the Eurasian Economic Union adds a further complication. Russia, Kazakhstan, Kyrgyzstan, Armenia and Belarus form a customs union in which goods move without customs controls at internal borders. This affects how trade between members is recorded, and it is one reason why figures for the same flow can differ. Membership questions have been under discussion, so the position for the period under review should be checked.

Working with product codes and priority lists

Trade data is organised by the Harmonized System, the product nomenclature maintained by the World Customs Organization. Codes are harmonised internationally to six digits, and many countries add further digits for their own purposes. Analysis at six digits allows comparison between countries. National codes can be more precise but may not match across borders.

The EU, the UK, the US and Japan have jointly developed a common list of high priority items. It groups by HS code the goods identified as used in Russian military systems or as important to their production, many of them electronic components. The list gives researchers a defensible starting point. It also has limits. A single HS code can cover many products, most of them in ordinary civilian use, and the list is updated from time to time, so the version used should be recorded.

From patterns to companies

Aggregated statistics cannot identify individual companies. Moving from a national pattern to a named business requires other sources.

• Company registers. Registration data and its history can show when a trading company was formed, who owns and manages it, and whether its stated activity changed after 2022.
• Corporate histories. Earlier names, addresses and directors can link a new entity to an established business in Russia or elsewhere.
• Transport records. Shipping and logistics data can add detail where it is lawfully available, although much overland trade in the region moves by road and rail and leaves a thinner public record than maritime cargo.
• Commercial trade databases. Some providers sell transaction-level customs data, but coverage varies widely and the origin of the data should be checked, because records obtained unlawfully create legal and reputational risk for anyone who uses them.

Consider a hypothetical. Suppose a compliance team finds that a recently formed trading company in a neighbouring state exports electronic components to Russia, and that its sole director previously managed a Russian distributor of the same goods. That is a meaningful finding. It is not, on its own, proof of sanctions evasion.

Limits and grading

Anomalies in trade data have many innocent explanations. Exporters and importers value goods differently, record them at different times and sometimes classify the same product under different codes. Goods may be stored, processed or consumed locally. Genuine new demand in a growing economy can look like diversion. Statistical revisions can change a picture after it has been published.

A pattern in national trade statistics can justify a question about a company, and it cannot by itself answer one.

For this reason, inference should be graded. We separate what a source states, what can reasonably be inferred from it and what remains open, and we record the confidence attached to each step. A finding that a company's trade is consistent with re-export differs from a finding that it re-exported a specific controlled item, and a report should say which one the evidence supports.

Sanctions compliance and legal advice

Research of this kind usually feeds decisions about sanctions and export control compliance. Those decisions turn on legal questions that research cannot settle, including which regimes apply, how ownership and control rules such as the OFAC 50 Percent Rule operate, and what a business knew or should have known. Rules in this area change often, and the position should be checked at the time of research.

Our approach is to use lawful sources, show our working and state plainly where the evidence stops. A well-documented open-source picture helps counsel and compliance teams decide where to focus their own enquiries. It is most useful when candid about its limits.