Asia Compliance Forum Export Control Working Group
Research report | China | January–June 2026 decision dates
China’s export-control enforcement is best understood through the operational decisions that precede shipment: how a product is classified, which technical specifications are verified, whether a licence covers the transaction, and what is declared to customs. A review of 320 administrative decisions dated in the first half of 2026 shows these issues recurring across a broad range of goods and customs authorities.
The sample contains RMB 27.60 million in disclosed fines and decisions issued by 100 customs authorities. Licensing-related conduct is the most frequently identified theme. Industrial equipment and metal materials, alongside rare earths and permanent magnets, account for a substantial part of the item coverage. Outcomes range from no penalty and warnings to fines and confiscation.
These findings describe a curated document sample, not the national enforcement total. They support an operational reading of the available decisions; they do not establish the probability that a shipment will be inspected or penalised.
1. The enforcement picture in numbers
320 decisions dated 1 January–30 June 2026, issued by 100 customs authorities.
292 decisions with positive disclosed fines, representing 91.3% of the sample.
RMB 27,597,667.26 in disclosed fines, a lower bound because one decision redacts an additional individual fine.
RMB 28,350 median disclosed fine among decisions with positive fines.
27 no-penalty decisions and one warning decision.
Of the 292 decisions with fines, 258 are classified as fine-only outcomes and 34 as fines accompanied by confiscation. The no-penalty count refers to decisions classified as such overall; it excludes documents that waive a penalty for only part of the conduct while imposing a fine for another violation.
2. Monthly volume is uneven; it does not prove a change in enforcement intensity
The sample contains 56 January decisions, 48 in February, 72 in March, 58 in April, 52 in May and 34 in June. March is the largest month, accounting for 22.5% of the sample. June is the smallest.
Decision counts and fine totals do not move together. March has the highest decision count, while April has the largest disclosed monthly fine total, approximately RMB 5.77 million. June’s 34 decisions carry approximately RMB 4.33 million in disclosed fines, compared with approximately RMB 3.98 million across March’s 72 decisions.
The evidence therefore does not support treating a lower monthly count as lower financial exposure. Nor does it establish a national slowdown or acceleration. Collection coverage, case complexity and the interval between conduct and a decision can all affect the observed pattern. A first-half decision may concern activity that occurred before 2026.
3. Licensing and declaration accuracy are connected enforcement issues
A licensing-related tag—failure to obtain or submit a licence—appears in 276 decisions, or 86.3% of the sample. Inaccurate, false or concealed declaration wording appears in 99 decisions, or 30.9%. Four decisions carry an outside-licence-scope tag, and 21 contain wording classified as a smuggling finding.
These are overlapping, keyword-assisted analytical categories. They are not mutually exclusive legal classifications. In particular, the licensing category combines different situations: failing to obtain authorisation and failing to submit the required licence are not necessarily the same conduct. The smuggling tag does not establish a criminal conviction.
The recurring connection between licensing and declarations suggests that export review should be assessed as a connected process. A product classification has limited practical value if the result is not reflected in the licence check, shipping instructions and customs declaration. Conversely, a declaration that repeats a commercial description may not communicate the technical characteristics relevant to control status.
4. Exposure extends across materials, equipment and specialised goods
The largest item theme is industrial equipment and metal materials, appearing in 114 decisions, followed by rare earths and permanent magnets in 90. Other themes include graphite in 34, tungsten/molybdenum/antimony/bismuth/tellurium in 26, chemicals and batteries in 24, and gallium and germanium in 22. Aviation, drones and navigation appear in 12 decisions; military items and protective equipment in ten; carbon fibre and glass fibre in two each.
The categories overlap and differ in breadth. The 114-decision equipment-and-metals category is wider than a single-material category such as graphite. These figures should not be read as comparable industry enforcement rates, and a material name alone does not establish whether a particular product is controlled.
Our reading is that product-level evidence deserves greater attention than broad sector labels. Composition, performance, specifications and the configuration of the exported item can be more informative than the industry in which the exporter operates. Where a transaction depends on a supplier’s technical statement, the useful question is whether that statement is specific enough to support the actual item and shipment being reviewed.
5. Enforcement is geographically distributed
Shanghai Pudong International Airport Customs issued 33 decisions in the sample, Shanghai Waigaoqiao Port Area Customs 30, Shekou Customs 23, Beilun Customs 17 and Huangdao Customs 16. Together, these five authorities account for 119 decisions, or 37.2% of the sample. The remaining decisions are distributed across 95 authorities.
This distribution demonstrates breadth within the collected records. It does not rank customs districts by inspection intensity. The issuing authority is not necessarily the exporter’s location or the port where goods were intercepted, and the dataset contains no trade-volume denominator with which to calculate comparative enforcement rates.
Detection wording also spans different processes. Inspection or testing is identified in 148 decisions, query or identification in 42, and audit in 14. These tags may overlap and are not a complete reconstruction of each investigation. Their practical significance is that relevant evidence can arise through technical examination or subsequent review, as well as at the point of declaration.
6. The fine distribution is more informative than the aggregate alone
The median positive disclosed fine is RMB 28,350, substantially below several of the largest amounts. The ten largest decision-level disclosed fines total RMB 7.902 million—approximately 28.6% of all disclosed fines in the sample. The largest is RMB 1.39 million.
The aggregate is therefore influenced by a relatively small upper tail. It would be misleading to use the average or the largest fine as a prediction for an individual transaction. Differences in conduct, goods, transaction value, the number of parties and the treatment of other violations all complicate such comparisons.
The monetary measure also has defined boundaries. It sums disclosed fines at decision level, including separately imposed fines on multiple parties where present. It excludes confiscated proceeds, deposits and late-payment additions. Some documents address other declaration violations alongside export-control conduct, so the total cannot be described as export-control-only fines. Redacted amounts are not estimated.
7. Penalty considerations require careful interpretation
Admission and acceptance of penalty appear as an extracted consideration in 210 decisions. Mitigation below the statutory range appears in 87 and a lighter penalty within the statutory range in six. One decision carries an explicit voluntary-disclosure tag. These categories overlap.
The frequency of a phrase is not evidence of its independent effect on the outcome. The sample cannot establish what penalty would have been imposed without cooperation, nor does a single voluntary-disclosure tag mean that only one exporter cooperated voluntarily. Wording and the detail available in each document vary.
The 27 no-penalty decisions likewise should not be treated as a general exemption. Their existence shows variation in outcomes within the sample. It does not establish that similar goods, low transaction values or subsequent corrective action will produce the same treatment in another matter.
8. Questions for an export-control programme
Viewed together, the decisions support five practical questions for internal review:
Classification evidence: Is there a documented basis for the classification of the actual product, including the relevant technical characteristics?
Licence alignment: Is authorisation, where applicable, checked against the goods and transaction being shipped rather than treated as a general approval?
Declaration consistency: Do the customs declaration, invoice, specifications and internal classification record describe the same item consistently?
Evidence retention: Can the business reconstruct the information available when a shipment was approved, including supplier inputs and subsequent changes?
Escalation quality: Do uncertainty, discrepancies and regulator queries reach people who can assess the technical and transaction facts before decisions are made?
These are management questions drawn from observed patterns, not a transaction-specific legal determination. The central lesson from this sample is the importance of connecting technical knowledge, licensing decisions and customs execution. A policy can set the standard; the shipment record shows whether that standard was applied.
About the research
This report uses only the supplied 2026 document collection. Of 378 PDFs reviewed, 320 unique decisions form the core first-half sample. Forty-eight decisions fall in the second half, four files duplicate decision numbers, and six records remain outside the core sample pending date or identifying-information verification. The body decision date takes priority over a conflicting header date. The unit is a unique decision number, not an individual company or shipment.
The collection includes export-control, military-item, relevant dual-use licensing and controlled-item declaration matters, including warnings and no-penalty decisions. It is not a census of Chinese enforcement or criminal proceedings. Item, conduct, detection and penalty-consideration tags are extracted from document wording and may overlap. OCR and extraction limitations remain; missing information is not treated as zero or as proof of absence. The dataset does not provide complete destination, end-user, payment or appeal-outcome coverage.
Explore the accompanying H1 2026 interactive enforcement analysis in Resources.
