The quietest channel in financial services
Search most anti-money-laundering training for the words life insurance and you will find nothing. The sector is treated as a cousin of banking that happens to sell protection, and its own language is actuarial: mortality, lapse rates, persistency, reserving. None of that vocabulary contains a criminal.
What a life or investment-related policy actually is, stripped of the sales description, is a long-dated store of value with four properties that matter here:
- it can be bought with one large payment and nothing more;
- it can usually be surrendered for cash, sometimes early and at a penalty;
- it can frequently be assigned to somebody else, so the value changes hands without a payment;
- and it can commonly be used as collateral for borrowing, so its value can be realised without selling it.
Put those together and a single-premium policy is a bearer-like asset with a regulated institution's name on it, a contract note, a tax identity and a long paper trail of legitimacy.
And the person who eventually receives the money may be unknown for decades. That is not an oversight in the controls; it is how the product works, and European law says so explicitly, which lesson three sets out. No other financial product tolerates a counterparty whose identity is established only at the moment of payment.
Why this is also structurally hard to police. The policy is usually sold by an intermediary, the premium often arrives from somebody's bank rather than from the insurer's own customer, the relationship may last thirty years with no contact, and the claim is handled by a team measured on how quickly it pays a grieving family. Each of those is reasonable. Together they describe a channel in which identity is collected once by somebody else, value accrues invisibly, and the control question arrives at the worst possible moment.
What this course does. It takes the European instruments as the worked example, because they are the clearest public statement of the duties, and reads the risk out of them rather than from a list of typologies. The pattern is not European: wherever policies can be bought with single premiums, surrendered, assigned or pledged, the same arrangement works.
A note on tone. Nothing here suggests insurers are negligent. The sector's controls are usually sound at inception and thin exactly where this product is unusual, which is everything that happens after inception.
Who is in scope, and why the label does not decide
Scope in this sector is easy to get wrong, because the obvious answer, the insurer, is only part of it.
Under Directive (EU) 2015/849, the definition of a financial institution includes an insurance intermediary where it acts with respect to life insurance and other investment-related services, with the exception of a tied insurance intermediary. So a broker arranging an investment-linked policy is an obliged entity in its own right; the same broker arranging motor insurance is not acting in that capacity at all. It is the activity that brings the firm into scope, not its trade description.
Under Regulation (EU) 2024/1624, which applies from 10 July 2027, the scope is drawn more carefully and the test changes in an instructive way. It covers insurance undertakings, and expressly includes insurance holding companies and mixed-activity insurance holding companies as defined in the Solvency II Directive. For intermediaries, it covers those acting with respect to life insurance and other investment-related insurance services, with an exception for an intermediary that does not collect premiums or amounts intended for the customer and which acts under the responsibility of one or more insurance undertakings or intermediaries for the products concerned.
Read that exception closely, because it is the better test. The older formulation turned on whether the intermediary was tied, which is a question about contracts and branding. The newer one turns on whether the intermediary handles the money. An intermediary through whose hands premiums or payouts pass is in scope; one that merely introduces business under an insurer's responsibility is not.
That is the right question, and it is worth adopting regardless of which instrument binds you: follow the money, not the label. Ask of each party in the chain whether client money touches it. If it does, it needs controls of its own, whatever the distribution agreement calls it.
The practical consequences.
An insurer may be relying on due diligence performed by a firm that is itself an obliged entity, or by one that is not, and the distribution agreement rarely makes the difference visible to the compliance function.
Group structures matter: bringing holding companies expressly into scope closes an argument about where the obligation sits when the regulated carrier is a subsidiary.
And a firm that sells both protection and investment products is in scope for part of what it does, which means its controls have to be able to tell its own business lines apart.
A counterparty you are allowed not to know
Here is the provision that makes this product unlike anything else, and it deserves quoting rather than summarising.
Article 13(5) of Directive (EU) 2015/849 requires that, for life or other investment-related insurance business, in addition to the customer due diligence (CDD, which much of the industry calls KYC) measures required for the customer and the beneficial owner, institutions conduct customer due diligence on the beneficiaries of the policy, as soon as the beneficiaries are identified or designated:
- where beneficiaries are identified as specifically named persons or legal arrangements, by taking the name of the person; and
- where beneficiaries are designated by characteristics or by class or by other means, by obtaining sufficient information concerning those beneficiaries to satisfy the institution that it will be able to establish the identity of the beneficiary at the time of the payout, or at the time of the exercise by the beneficiary of its vested rights.
Three things follow, and each is unusual.
First, the beneficiary is a third due diligence subject. Not the customer, not the beneficial owner of the customer, but a further person: whoever will be paid. Most institutions' systems have two slots and this product needs three.
Second, for a named beneficiary the requirement at that stage is to take the name. Not to verify, not to screen in the ordinary sense: to take the name, as soon as the person is identified or designated. The verification obligation attaches later.
Third, and most striking, the law expressly contemplates not knowing. Where the beneficiary is a class, such as my children living at my death, the duty is to hold enough information to be able to establish identity when the money is paid. The identification is lawfully deferred, sometimes for decades.
Why that is a control problem rather than a loophole. It means the decisive control cannot sit at onboarding, because at onboarding there may be nobody to identify. It has to sit at payout, or at the moment vested rights are exercised. And payout is operationally the worst moment available: a death claim, a bereaved family, a service target measured in days, and a claims team whose training is in sympathy and settlement rather than in financial crime.
The honest summary. Every other part of this catalogue says identify before you transact. This product says you may not be able to, so be ready to identify at the end. Institutions that treat the claim as an administrative step have put their only effective control in the hands of the function least equipped to exercise it.
Reading the risk out of the lower-risk list
Annex II of the Directive lists factors that may indicate lower risk. Three of them concern this sector, and they are more useful read backwards, because each tells you precisely which feature of a policy carries the risk.
The Annex names as potentially lower risk:
- life insurance policies for which the premium is low;
- insurance policies for pension schemes if there is no early surrender option and the policy cannot be used as collateral; and
- a pension, superannuation or similar scheme providing retirement benefits to employees, where contributions are made by way of deduction from wages, and the scheme rules do not permit the assignment of a member's interest under the scheme.
Now invert each clause. The law is telling you that risk rises with:
A large or single premium. The lower-risk case is a low premium, so the concern is a policy that absorbs a lot of money at once. A single-premium bond is placement in one transaction.
An early surrender option. The lower-risk case has none. Surrender is the exit: money in, money out, with the policy as the intermediate owner of the value. Where surrender is possible early, the holding period can be short.
Use as collateral. The lower-risk case cannot be pledged. Where a policy can be, its value is realisable as a loan drawdown, which arrives in a bank account as borrowing rather than as insurance proceeds and looks entirely different in monitoring.
Assignability. The lower-risk case forbids assignment of a member's interest. Where assignment is permitted, the policy's value transfers to another person with no payment at all, which is the same property this catalogue's course on free-of-payment transfers describes in securities.
Contributions that are not wage deductions. The lower-risk case is funded straight from payroll, which is to say from a source already documented by an employer. Any other funding route is a source-of-funds question.
Why this matters more than a typology list. A list of typologies goes out of date. These five features are structural: they are the product design choices that make a policy useful for laundering, and they appear in a product's terms and conditions where anybody can read them. An institution can classify its own book by them, today, without any new data.
How the methods actually work
With the five features from the previous lesson in hand, the methods are almost predictable. Each uses one of them.
Single-premium placement and early surrender. A large premium is paid into an investment-linked policy. Some months later the policy is surrendered, accepting the penalty. The customer receives a payment from a regulated insurer, with a contract, a policy number and a tax statement. The penalty is the cost of laundering, and a surrender loss accepted without complaint is itself the signal: ordinary customers argue about charges.
Overpayment and refund. More is paid than the premium due, and the excess is refunded, sometimes to a different account or a different person. The refund is a payment from an insurer, which is about as unremarkable as a credit can be.
Assignment for value. The policy is assigned to a third party, who later surrenders or claims. Value has moved to someone the insurer never underwrote and may never have identified, and no payment accompanied the transfer.
Policy as loan collateral. The policy is pledged, a loan is drawn, and the proceeds reach a bank account as borrowing. If the loan is never repaid, the policy settles it, and the money has been extracted without a surrender ever appearing.
Third-party premium payers. Premiums arrive from somebody other than the policyholder: a relative, an employer, a company, an account in another country. Each has innocent explanations, and together with the others on this list they rarely do.
Free-look cancellation. Many markets give a short cancellation window with a full refund. Money in and money straight back out, with the insurer as a one-week custodian.
Changing the beneficiary late. The beneficiary is altered shortly before a claim, or shortly after the policyholder's circumstances change, so the person paid is not the person contemplated when any due diligence was done.
The investment wrapper as a custody account. An investment-linked policy that holds assets chosen by the customer is, functionally, a portfolio inside an insurance contract, and a transfer of the policy transfers the portfolio.
What they have in common. In each case the money leaves as an insurance payment: a surrender, a refund, a claim or a loan drawdown. That is the point. The laundering is not hidden inside the insurance; the insurance is what makes it look ordinary afterwards.
The distribution chain, and who holds the file
Almost nothing in this sector is sold by the institution that carries the risk, and that single fact explains most of the control gaps.
The usual chain. A customer deals with a broker, an agent, a bank branch selling insurance, or a tied representative. That party gathers the identity documents, explains the product and submits the application. The insurer underwrites, issues the policy and holds the liability. The premium may pass through the intermediary's client account or go directly to the insurer. Thirty years later the insurer pays a claim to a person it has never met, on the strength of a file assembled by a firm it may no longer deal with.
Where the due diligence actually sits. With whoever met the customer. The insurer's own record is frequently a set of copies, or a declaration that the intermediary performed the checks, or nothing beyond the application form. The lawful position may be perfectly in order and the practical position is that the institution bearing the obligation has not read the evidence.
Which is why the money test from lesson two matters. The newer European formulation asks whether the intermediary collects premiums or amounts intended for the customer. Apply it to your own chain and you will usually find three categories:
- intermediaries that handle client money, which are obliged entities with their own duties, their own supervisor and their own failures;
- intermediaries that only introduce, under the insurer's responsibility, where the insurer's controls are the only controls; and
- arrangements nobody has classified, which is the largest category in most insurers.
Two specific weaknesses to look for.
Reliance without inspection. Where an insurer relies on an intermediary's due diligence, somebody should have tested a sample of those files. Reliance that has never been sampled is an assumption.
Orphaned policies. When an intermediary ceases to act, its book transfers or lapses into the insurer's direct administration. The customer relationship ends and the policy continues, which means the file is now the only knowledge that exists, and it was collected by a firm that is gone.
The uncomfortable question for an insurer. For a policy written twenty years ago through an intermediary no longer on your panel, who in your organisation could say today who the customer was, where the premium came from and who is designated to be paid? If the honest answer is nobody, that is not a historical problem: it is the state of the control at the moment a claim arrives.
What to watch, and when
Because the product has a thirty-year life and three different counterparties, monitoring has to be organised by event rather than by transaction. These are the events worth a question, in the order they occur.
At inception.
- Premium size against the customer's stated means and occupation, with source of wealth rather than source of funds for a large single premium.
- A third-party premium payer, which should be identified and explained, not merely recorded.
- A product chosen for features the customer cannot explain, particularly early surrender or assignability.
- A policy whose term, cover or structure makes no sense for the customer's stated need, such as large cover for someone with no dependants.
During the life of the policy.
- Changes of beneficiary, and especially changes to a beneficiary with no apparent relationship to the policyholder.
- Assignment of the policy, which is a change of economic owner.
- A request to use the policy as collateral, and who the lender is.
- Top-ups out of proportion to the original premium.
- Changes of address, bank details or correspondence to a different country, which is the same control problem as payment diversion.
At surrender.
- Early surrender where a penalty is accepted without complaint.
- Surrender proceeds paid to an account or a person other than the policyholder.
- Surrender shortly after a top-up, which is money in and straight out with a policy wrapped around it.
At claim or payout.
- Identification and verification of the beneficiary, which is the obligation the law defers to this point.
- The beneficiary's relationship to the policyholder, which is the question least often asked and most often informative.
- A beneficiary that is a company, trust or other legal arrangement rather than a person.
- Screening at payout, not only at inception, since sanctions and political exposure change over thirty years.
The honest obstacle. Most insurance systems are policy-centric: they answer questions about a policy and struggle to answer questions about a person. Ask whether your systems can show every policy on which one individual appears as policyholder, premium payer, assignee or beneficiary. Where they cannot, none of the patterns above can be seen across a book, only inside one contract at a time.
The claim, which is where the control lives
Everything in this course converges on one moment. The law defers identification of the beneficiary to the payout or the exercise of vested rights, so the payout is where the only complete control can be exercised. It is worth treating that moment deliberately, because everything about it works against doing so.
What makes it difficult. A death claim arrives with a grieving family, a certificate, and an expectation of speed. Claims functions are measured on settlement times and complaint volumes, and rightly so: delay in paying a widow is a real harm, not a compliance inconvenience. Asking a bereaved person to evidence their relationship to the deceased feels, and can be, callous.
So design the control to respect that rather than fight it.
- Ask early. Where the beneficiary is known during the policy's life, take the name and keep it current, as the law requires as soon as the beneficiary is identified or designated. Every question answered before the claim is a question not asked at the worst time.
- Separate the decision from the conversation. The person speaking to the family should not be the person assessing financial crime risk, and should not have to.
- Set out in advance what evidence is proportionate for a claim of a given size, so that nobody is improvising a demand at the counter.
- Screen at payout as a matter of course, because thirty years have passed since anyone last looked, and the beneficiary has never been screened at all.
- Treat a beneficiary that is a legal arrangement as a different case, with its own beneficial ownership question, rather than as a name on a form.
- Record the relationship, which is the single most useful fact and takes one question.
And know what a late change means. A beneficiary altered shortly before death, or to someone with no evident connection, is the one pattern this product offers that has no innocent reading in volume. It will usually be innocent in the individual case, which is precisely why it has to be reviewed rather than assumed either way.
The governance point. If an institution's only effective anti-money-laundering control sits inside its claims process, then its claims process is part of its control framework, and should be staffed, trained, measured and audited as such. In most insurers it is none of those things, because nobody has noticed that this is where the control ended up.
Building it, and the limits
A programme for this sector is unglamorous and mostly consists of asking questions at moments nobody currently asks them.
What to put in place.
- Classify your own book by the five features: single or large premium, early surrender available, assignable, pledgeable, and funded otherwise than by wage deduction. That is a product-terms exercise, not a data project, and it tells you which policies deserve attention.
- A third due diligence subject. Make the beneficiary a first-class record alongside the customer and the beneficial owner, captured as soon as identified or designated, as the Directive requires.
- Event-based monitoring on the list in lesson seven, rather than transaction monitoring borrowed from banking.
- A person-centric view, so one individual can be found across policyholder, payer, assignee and beneficiary roles.
- Sampling of intermediaries' files, so that reliance is tested rather than assumed.
- A claims control, designed as set out in the previous lesson and owned by somebody.
- Screening at payout, every time.
What to measure. The share of in-force policies with a beneficiary recorded at all. The share of large single premiums with source of wealth evidenced. The number of policies where the premium payer differs from the policyholder, and how many of those payers are identified. Time to complete beneficiary verification at claim, which tells you whether the control is real or waived under pressure. And the proportion of intermediary files sampled in the last year.
The limits, honestly.
The insurer has one relationship and the money frequently arrives from, and leaves to, parties it has never met. No amount of diligence on the policyholder changes that.
The law permits the beneficiary to be unknown until payout, so for a class designation there is nothing to verify for years, and pretending otherwise produces false comfort.
Policy administration systems were built to administer policies, and the person-centric view is a genuine project rather than a configuration change.
And a thirty-year product outlives the intermediary, the system, the staff and often the control standard that applied when it was written.
The sentence to carry away. In banking, you know the counterparty and watch the transactions. In life insurance, there are few transactions, and you may not know the counterparty until the day you pay them. The control has to be built for that, and almost none are.
Thirty questions
Answer all thirty questions, then check your answers. You need 27 out of 30 to pass and receive a certificate. If you score less, you will see which answers were right and wrong, and then go through the course again before you retake the check. Your answers, progress and times are kept only in this browser.
Your answers
Your certificate of completion
Enter your name as you want it to appear, then save the certificate as a PDF. In the print window, choose Save as PDF. A certificate is issued once per completion of the course.
Rate this course
How would you rate this course? Your rating is anonymous — just a number, with no name and no account.
Bring this to your team
Your colleagues can take this course too — it is free, needs no sign-up, and nothing they enter leaves their browser. If your institution wants the same ground covered as structured training for a team, with an instructor and your own cases, we run courses in a classroom or online for groups of ten or more.
Talk to us about training your team → See the paid courses →
Saolix does not record who takes this course, so it cannot verify these certificates. The certificate confirms completion of a free self-paced course and is not an accredited qualification.
Sources
The official documents this course relies on. Laws and guidance change, so check the current version.
- Directive (EU) 2015/849, Article 3(2), Article 13(5) and Annex II (lower-risk factors) · EUR-Lex
- Regulation (EU) 2024/1624 (the scope of insurance undertakings, holding companies and intermediaries) · EUR-Lex
- Directive 2009/138/EC (Solvency II), Article 212(1) (insurance holding companies) · EUR-Lex
- Directive (EU) 2016/97 on insurance distribution, Article 2(1) (insurance intermediary) · EUR-Lex



