A unit price test on an importer's shipments, the value that may have moved, and the mirror check against the exporter's own declaration.
Test the declared unit price of each shipment against a price band built from comparable trade, quantify the value above a reference price, then confirm it with the exporter's own declaration of the same goods. On an illustrative importer with 20 shipments, seven exceed the upper fence of 14.75 euros per kg around a 12.00 median, and 1,854,200 euros, 27.3 per cent of the value invoiced, sits above the median price. For the four shipments at 31.50, the mirror check against the exporter's FOB declaration puts the gap at 1,468,480.
Worked in full in The EU AML Handbook by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Trade-based money laundering moves value through the price of goods rather than through the payment. An importer that pays 31.50 a kilogram for goods worth about 12.00 sends the difference abroad with an invoice to justify it; an exporter that under-invoices does the reverse. The payment itself looks like any trade settlement, which is why payment monitoring misses most of these cases and why trade finance and correspondent banking teams test unit prices directly.
| Input | Value |
|---|---|
| Comparable imports in the reference set (same product code, origin, quarter) | 15 |
| Reference prices, lowest and highest | 9.70 and 14.60 |
| Customer shipments tested | 20 |
| Customer volume and invoiced value | 401,000 kg; 6,780,100 |
| Exporter's declared FOB price, flagged shipments | 12.40 |
| Freight and insurance, CIF over FOB | 6% |
The reference set excludes the customer and should be as narrow as the data allows: the most detailed product code available, the same origin and period, similar lot sizes. Quality grades within one code can legitimately differ a great deal, which is the main source of false positives in price testing.
Use quartiles rather than the mean: they are not moved by the very anomalies you are looking for, and trade prices are rarely symmetric.
Q1 = 11.50, median = 12.00, Q3 = 12.80; interquartile range (IQR) = 1.30
Upper fence = Q3 + 1.5 × IQR = 14.75; lower fence = Q1 − 1.5 × IQR = 9.55
In Excel, with reference prices in A2:A16: =QUARTILE.INC(A2:A16,3)+1.5*(QUARTILE.INC(A2:A16,3)-QUARTILE.INC(A2:A16,1)). A price above the upper fence suggests over-invoicing on an import; below the lower fence, under-invoicing.
| Shipment | Unit price | Ratio to median | Invoiced | Excess over median |
|---|---|---|---|---|
| 5, 10, 14, 17 (each) | 31.50 | 2.62 | 630,000 | 390,000 |
| 8 | 18.90 | 1.57 | 378,000 | 138,000 |
| 13 | 16.20 | 1.35 | 340,200 | 88,200 |
| 19 | 15.40 | 1.28 | 308,000 | 68,000 |
| Seven flags | 3,546,200 | 1,854,200 |
Excess value = Σ (declared price − reference price) × quantity
At the median as reference, 1,854,200; at the upper fence, the part no plausible quality difference explains, 1,466,450. Report both: the first is the best estimate of value moved, the second the floor. Four identical prices of 31.50 are a pattern in themselves: genuine trade prices drift, contrived ones repeat.
Every import is someone else's export, declared to another customs authority. Where the exporter's declaration or the supplier's documents are available, through trade data, a correspondent bank or a request to the customer, compare the two. The exporter declared the four 31.50 shipments at 12.40 FOB. Adding 6 per cent for freight and insurance gives an expected CIF price of 13.14.
Gap = (31.50 − 13.14) × 80,000 kg = 18.36 × 80,000 = 1,468,480
The goods left the origin country valued at 992,000 and arrived valued at 2,520,000. The difference is not explained by the goods, and it was paid.
| Rule | Upper limit | Flags | Excess over median |
|---|---|---|---|
| Q3 + 1.0 × IQR | 14.10 | 7 | 1,854,200 |
| Q3 + 1.5 × IQR | 14.75 | 7 | 1,854,200 |
| Q3 + 3.0 × IQR | 16.70 | 5 | 1,698,000 |
| Median + 25% | 15.00 | 7 | 1,854,200 |
The value at stake barely moves with the rule, because it sits in the four extreme shipments: the wide 3.0 fence still captures 1,698,000. What moves is the alert count, and with it analyst time. For a large book, run the wide fence for alerts and the narrow one to size exposure once a customer is under review.
Comparing prices on different delivery bases. Imports are usually valued CIF, including freight and insurance; exports FOB. Testing a CIF import price against an FOB reference makes every clean shipment look over-invoiced by the freight share, 6.0 per cent here, and a whole product line can alert. The reverse error, testing an FOB export price against CIF comparables, makes clean exports look under-invoiced and can mask modest over-invoicing. Put every price on one basis before building the band. The second mistake is ignoring quantity: a modest overprice on a large lot can move more value than an extreme price on a small one, so rank alerts by excess value, not by ratio.
A flag is a reason to ask, not a finding. Ask for the contract, the specification and grade, the shipping documents and the payment route; check whether the counterparty is related and whether the goods fit the customer's business. Mispricing combined with a third-party payer or routing through a free trade zone is a far stronger indicator than price alone.
How to measure what monitoring misses is worked in the detection rate calculation, and the customer-behaviour equivalent of a price band is in the peer group z-score article.
Build a narrow reference set on one price basis, flag prices beyond the quartile fences, size the excess at the median and at the fence, and confirm the largest cases against the exporter's declaration: here 1,854,200 euros of excess on 6,780,100 invoiced, and a 1,468,480 mirror gap on four shipments. The EU AML Handbook covers what the AMLR, applying from 10 July 2027, expects of risk-based monitoring, and the free workbook and working documents for the book size the remediation effort behind it.
A comparison of the importer's declared value with the exporter's declaration of the same goods, after converting both to one basis. In the illustrative case the exporter declared 12.40 FOB, about 13.14 CIF with 6 per cent freight and insurance, against 31.50 invoiced to the importer: a gap of 1,468,480 on 80,000 kg.
There is no fixed percentage; it depends on how dispersed prices are for the product. A common approach flags prices above the third quartile plus 1.5 times the interquartile range of comparable trades. With quartiles of 11.50 and 12.80 the fence is 14.75 euros per kg, 1.23 times the 12.00 median.
Yes, on exports. An exporter that under-invoices is paid less than the goods are worth, so the importer receives the excess value, often with the difference settled informally, moving value into the importer's country. The same test applies in mirror image: flag prices below the lower fence, 9.55 euros per kg in the illustrative reference set.
This article is one calculation from The EU AML Handbook. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: regulation, compliance and banking → · All 1682 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.