Correspondent banking and wire layering

As of September 2026: Danske Bank's US guilty plea was entered in December 2022, and the bank described a September 2024 French settlement as the final investigation by any authority into its Estonian non-resident portfolio. Rules such as the Section 312 and 313 requirements can be amended, so check current regulator text before relying on any detail here.

Also called: nested accounts · nesting · wire layering · downstream correspondent banking · payable-through accounts.

What is correspondent banking, and where does wire layering come in?

Most banks are small next to the payment systems they need. A bank in Tallinn or Riga cannot settle US dollars by itself, and neither can most banks anywhere. So it opens an account at a large bank in a dollar centre, which moves the money on its behalf. That arrangement is correspondent banking. The large bank is the correspondent, and the smaller bank that uses it is the respondent. It is a normal, essential part of world trade, and it is legal.

Wire layering is what a launderer does with that plumbing. A wire is a bank-to-bank transfer, and each one adds a line to the trail. Chain enough of them together, across enough accounts and countries, and the layering stage does its job: the money’s origin is far behind it, and every step needs a records request to a different bank or country.

The correspondent model has a built-in blind spot. The large bank knows its customer, the respondent bank. It does not see that bank’s clients, and it cannot vet them one by one. A related arrangement makes the blind spot deeper. In nesting, the respondent uses its account to serve other banks, so a third or fourth institution rides inside the relationship. Nesting is common and often legitimate, but when the top-level bank is not told about it, it is serving customers it never approved.

How does the technique work?

1. A criminal or corrupt insider needs a bank that will not ask hard questions. The bank is often small, located in a place with weak supervision, and eager for business from abroad.
2. The money owner sets up companies, frequently shell companies, in several jurisdictions. Their customers and owners are recorded on paper, but the real beneficial owner stays hidden.
3. The companies open accounts at the respondent bank. The bank onboards them with thin checks, sometimes with staff who help the customers keep the picture vague.
4. Wires move between the companies’ accounts, often inside the same bank first, then out to other banks in other countries. Descriptions are generic: loans, consulting, invoices.
5. Dollar payments pass through the correspondent bank’s account. To the correspondent, they look like the respondent bank’s ordinary business, a line in a large batch.
6. The funds land in third-party accounts abroad and are used for assets, spending, or another round of wires.

Why does correspondent banking work for launderers?

The model rests on trust between banks. A correspondent bank normally does not have a direct relationship with its respondent’s customers, so it must judge the respondent itself: who owns it, who supervises it, how good its controls are. If the respondent misstates those things, the correspondent is judging a picture that is partly false.

Speed and volume help too. Big correspondent banks handle enormous payment flows, and monitoring systems can only score what they can see. Payments bundled under one respondent’s name look like one customer’s activity.

Geography adds delay. Once funds have crossed two or three borders, an investigator needs legal requests to each country’s bank. Some countries answer slowly, and some do not answer at all. The structure does not have to be impossible to unravel, only slower than the case can afford.

Finally, the incentives are lopsided. A respondent bank earns fees from clients who pay to move money without questions. For a corrupt or careless bank, the risk is deferred and belongs partly to someone else.

Real case: Danske Bank Estonia

Danske Bank’s Estonian branch is the clearest recent example of the model under strain. According to the Justice Department, between 2008 and 2016 the branch offered a business line for non-resident customers, people and companies that lived outside Estonia, including in Russia. It attracted them by letting them move large sums with little oversight, and its employees worked with those customers to hide the true nature of their transactions, including by using shell companies that obscured who owned the funds.

The US link was the point. Those customers relied on US dollar payments, and US banks served as Danske’s correspondents. The Justice Department said the branch processed $160 billion through US banks for the non-resident customers. When those banks asked about the branch’s customers and anti-money laundering controls, Danske gave answers that prosecutors said were false.

On December 13, 2022, Danske Bank pleaded guilty to conspiracy to commit bank fraud and agreed to forfeit more than $2 billion. The charge is telling. The offence was not that the bank laundered money in the abstract. It was that the bank lied to its correspondents about its customers and controls. That is the trust mechanism above, broken from the inside. The full story, including the whistleblower, is on the cases page.

The Lebanese Canadian Bank shows the government’s response when the bank itself is the weak link. On February 10, 2011, the US Treasury named it a financial institution of primary money laundering concern under Section 311 of the USA PATRIOT Act. Treasury said a network run by a drug trafficker, Ayman Joumaa, laundered proceeds through the bank’s accounts and through Lebanese exchange houses, and proposed to bar US banks from holding correspondent accounts for it. A designation like that can effectively cut a bank off from the dollar system.

The same weakness appears when the risk comes from the customers’ side of the respondent’s business. The 2012 US Senate subcommittee report on HSBC described US dollar flows involving Mexican exchange houses suspected of laundering for cartels, moving through a large bank’s US operations. The details differ from Danske’s, but the lesson is the same: a bank inside the dollar system carries the risk of everything that reaches it, whether the customer is a respondent bank, an exchange house, or a company behind a shell.

How does it get caught?

The defence has three layers. The first is rules. Section 313 of the USA PATRIOT Act bars US banks from holding correspondent accounts for foreign shell banks that have no physical presence anywhere and are not regulated affiliates. Section 312 requires due diligence on foreign bank accounts and, for higher-risk banks, requires the US bank to ask about the respondent’s own owners and whether it serves other banks. The FATF’s Recommendations apply comparable expectations worldwide, and its 2016 guidance explains how to apply them without abandoning whole regions.

The second layer is the correspondent bank’s own monitoring. Investigators and compliance teams look at the respondent as a whole: volumes far above what its size suggests, a client base of non-residents, payments among related companies, and frequent use of round amounts. Warning signs also come from inside. A correspondent’s questions about a customer that go unanswered, or are answered with something that does not add up, are evidence in themselves.

The third layer is outside pressure: suspicious activity reports, whistleblowers, journalists, and government action. In Danske’s case an insider raised the alarm and the story became public through the bank’s own investigation and reporting that followed. See the detection overview for how these tools fit together.

The trade-off is real. Pressure on correspondent banks can lead them to close accounts across entire regions, which the FATF calls de-risking. That can cut off honest businesses and remittance senders and push activity out of sight. Good practice is to test each respondent on its own facts, and to treat a bank that hides its own customers as the risk, not the country it is in.

First published September 21, 2026 · Last updated September 21, 2026 · Legal statuses current as of September 2026