Eight Crime Policy Wording Mistakes That Leave CEO Fraud Uninsured
Most executive impersonation claims fail on policy wording, not on the fraud itself. Here are eight wording mistakes that leave a CEO fraud loss unpaid, and the clauses to ask for at your next renewal.
Why CEO fraud claims get declined even when the fraud is obvious
A controller gets an urgent email that looks like it came from the CEO, asking for a confidential wire to close an acquisition. The controller sends it. Two days later the real CEO has no idea what anyone is talking about, the money has moved through three accounts, and the company files a claim under its crime policy. Nobody disputes that a crime happened. The claim is declined anyway.
This is the most common way CEO fraud insurance coverage falls apart: the loss is real, but the wording doesn't reach it. Buyers usually find out after the money is gone. The eight mistakes below are ones brokers and coverage counsel commonly flag, and each comes with what to ask for instead. Wordings differ a lot by market, whether it's a London market crime form, a European fidelity policy, a US commercial crime form or a cyber endorsement sold in Singapore or Dubai. Treat this as a list of questions for your broker, not a legal opinion.
Mistake 1: Assuming computer fraud or funds transfer fraud covers executive impersonation
Most crime policies have a computer fraud or funds transfer fraud insuring agreement, and buyers often assume it covers a fake CEO email. It usually doesn't. Those clauses typically require unauthorized entry into your systems, or an instruction sent to your bank without your knowledge. In executive impersonation fraud, your own authorized employee knowingly sends the payment. They were deceived, but they acted on purpose. Courts in several jurisdictions have disagreed on how to read these clauses, and insurers have tightened the language in response.
Then there's the voluntary parting exclusion. It removes cover where someone with authority hands over money or property voluntarily, even if tricked into it. If your crime form carries that exclusion and has no separate social engineering or fraudulent instruction insuring agreement that expressly overrides it, assume a CEO fraud loss is excluded.
What to do instead: Confirm there's a standalone fraudulent instruction coverage or social engineering fraud insuring agreement. Then check that the voluntary parting exclusion is written not to apply to it. One line saying "this exclusion does not apply to the Fraudulent Instruction Insuring Agreement" is worth more than pages of marketing copy.
Verification conditions your finance team won't actually follow
Once the insuring agreement is sorted, the next trap is the verification condition. Many social engineering endorsements pay only if the employee checked the instruction first, usually by calling a number already on file rather than one given in the fraudulent message.
Mistake 2: Accepting a verification clause written as a condition precedent
Some wordings make verification a condition precedent to coverage. Skip the step and there's no claim, however convincing the fraud was. Others tie verification to a "documented procedure" the company must show it had and followed. Both look reasonable on paper. The problem is the gap between the wording and what happens on a Friday afternoon: a CEO who's traveling, a request marked confidential, and a finance team trained not to question the boss. That's the setup the fraudster builds.
What to do instead: Give the verification wording to whoever runs accounts payable and ask whether it matches what they do today. If it doesn't, fix the process or the wording before you bind. Where you can, negotiate verification as a retention or coinsurance trigger rather than a coverage killer, so a higher retention applies if verification was skipped instead of a full denial. Some insurers will agree for buyers with good controls. Others won't. Ask anyway.
Mistake 3: Letting the definition of who can be impersonated stay narrow
Check whose identity the fraudster has to fake. Some endorsements cover only impersonation of a vendor or client. Others cover only an "employee," defined so narrowly that directors, board members or a parent company's executives fall outside it. If a holding company sits in one country and subsidiaries operate elsewhere, a fake email from the group CFO may not count as employee impersonation for the subsidiary that paid.
What to do instead: Ask that the impersonated party include employees, directors, officers, partners and trustees of any insured entity, plus vendors, clients and other business relationships. Also check that "employee" covers people not formally on payroll in each country where you operate, such as seconded staff or contractors working inside your finance team.
Social engineering sublimits and retentions that don't match your payment sizes
A policy can cover CEO fraud in principle and still pay far less than you lost. The social engineering sublimit is where buyers get the biggest surprise.
Mistake 4: Buying a sublimit sized for small invoices, not for treasury transactions
A crime policy with a USD 5 million aggregate limit may carry a social engineering sublimit of USD 100,000 or USD 250,000. That suits a fake supplier changing bank details on a routine invoice. It does little for a fake acquisition wire, the typical shape of executive impersonation fraud. The sublimit sits inside the main limit. It isn't extra money on top.
What to do instead: Start from your largest single outbound payment that one or two approvers can release. That's your realistic worst case. Buy a sublimit that covers a meaningful share of it, knowing insurers will want to see payment controls before offering more. The ranges below are typical 2026 market figures only. Pricing varies by contract, region and the insurer's view of your controls, so confirm with your broker.
| Social engineering sublimit (USD) | Where it usually comes from | Indicative additional annual premium (USD) | What insurers usually ask for |
|---|---|---|---|
| 50,000 to 250,000 | Cyber policy endorsement, small-business crime forms | Often bundled; a few hundred to a few thousand if itemized | Short questionnaire on callback and dual approval |
| 250,000 to 1 million | Standalone crime or fidelity policy | Low thousands to mid-teens of thousands | Written payment procedures, bank-detail change controls |
| 1 million to 5 million and above | Crime policy with specialist underwriting, sometimes in layers | Roughly 10,000 to 60,000 or more | Detailed controls review, treasury segregation, training records |
Mistake 5: Ignoring a separate, higher retention for social engineering
Many wordings attach a separate retention to the social engineering agreement, often higher than the main crime retention. Hypothetically, say a 300-person manufacturer has a USD 250,000 sublimit and a USD 50,000 social engineering retention. A USD 400,000 fake-CEO wire then recovers USD 200,000 at most, depending on how the retention interacts with the sublimit. That's half the loss, before anything clawed back through the bank.
What to do instead: Run that arithmetic on your largest realistic payment before you renew. If the net recovery wouldn't change a board conversation, the cover is mostly cosmetic. Your options are a larger sublimit, a coinsurance share in exchange for a lower retention, or a better-sized crime policy.
Gaps between crime and cyber policies that leave the loss with no home
Plenty of companies hold both a crime policy and a cyber policy and assume one will respond. In practice the two can point at each other, and sometimes neither pays.
Mistake 6: Letting both policies exclude the loss, or both claim to be excess
Cyber policies often exclude transferred funds unless a social engineering endorsement is attached, and that endorsement is usually small. Crime policies may exclude losses "arising from" a network security event, which is easy to argue when the fake CEO email came from a compromised real mailbox rather than a lookalike domain. Then come the "other insurance" clauses. If both policies say they're excess of any other valid insurance, expect a long dispute and a slow payout.
What to do instead: Put both wordings side by side and decide which policy is primary for funds transfer loss. Usually that's crime, because the limits are bigger. Have the cyber policy pick up incident response costs: forensics, legal advice, and notification if the compromised mailbox exposed personal data under laws such as the GDPR, India's DPDP Act or Brazil's LGPD. Ask your broker to align the other-insurance clauses, ideally with the same insurer or a written priority-of-payment agreement.
Mistake 7: Overlooking ownership, direct loss and discovery wording
Three smaller definitions cause outsized trouble. First, ownership: if the stolen money belonged to a client, sat in escrow, or was held by a sister company outside the named insureds, the policy may treat it as not your loss. Second, direct loss: some forms pay only the immediate loss and exclude investigation costs, interest and contractual penalties. Third, discovery and notice: crime policies are often written on a discovery basis with strict notice periods. If you report late because you were busy chasing funds through the bank, you may hand the insurer a defense.
What to do instead: Ask for property "held by the insured in any capacity" to be covered, and confirm every group entity in every country is a named insured. Negotiate a reasonable sublimit for claim-preparation or investigation costs. Write the notice requirement into your incident playbook so a suspected impersonation payment triggers a call to the broker the same day as the call to the bank.
The renewal wording checklist to send your broker
Mistake 8: Approving the renewal without testing it against a real scenario
Most renewals get checked on premium, limit and retention, and that's it. The only reliable way to know whether you have CEO fraud insurance coverage is to put a specific scenario to the wording and ask the insurer, in writing, whether it would respond.
What to do instead: Write two short scenarios based on how your company really pays: a fake executive email ordering an urgent wire, and a fake supplier changing bank details. Send both to your broker 60 to 90 days before renewal with the requests below, and ask for answers in the binder or quote, not on a phone call.
- A standalone social engineering or fraudulent instruction insuring agreement, with the voluntary parting exclusion expressly carved back for it.
- Verification wording that matches your documented procedure, ideally triggering a higher retention instead of a denial.
- An impersonated-party definition covering directors, officers and employees of every insured entity, plus vendors and clients.
- A social engineering sublimit sized against your largest single payment that a small number of approvers can release, and a retention you've modeled.
- Clear priority between crime and cyber policies for funds transfer loss, with aligned other-insurance clauses.
- Ownership wording covering third-party funds held in any capacity, and all subsidiaries named across every country.
- Notice periods and claim-preparation cost cover that fit your incident playbook.
Coverage interpretation, insurance regulation and what local insurers can offer vary by country. A clause that's standard with a London market crime underwriter may not be available from a domestic insurer in Indonesia or Mexico. Before binding, have a licensed broker or coverage lawyer in your main jurisdiction review the final wording.
The decision rule is simple. If your broker can't confirm in writing that the fake-CEO scenario is covered, and what the net recovery would be, treat that exposure as uninsured and fund your payment controls on that assumption until the wording changes.
Quick answers
Does a standard crime policy cover CEO fraud?
Often it doesn't, unless the policy includes a specific social engineering or fraudulent instruction insuring agreement. Standard computer fraud and funds transfer fraud clauses usually need unauthorized system access or an instruction sent without your knowledge. The voluntary parting exclusion can also remove cover when an authorized employee was tricked into paying. Check for an express carve-back of that exclusion.
Should CEO fraud be covered under my cyber policy or my crime policy?
For most mid-size and large companies, the crime policy is the better primary home for the stolen funds, because it can carry larger sublimits. The cyber policy should respond to forensics, legal and data-breach costs if a mailbox was compromised. What matters is that both wordings agree on which policy is primary, so the loss doesn't fall between them.
How large should our social engineering sublimit be?
Start with the largest single outbound payment that one or two people can release, then decide how much of that loss you can absorb. A USD 100,000 sublimit may suit a small business paying routine invoices but leaves a treasury-sized wire mostly uninsured. Higher sublimits usually bring a closer look at your controls, and pricing varies by insurer and region, so get quotes at several levels.
What happens if our employee skipped the callback verification?
It depends on the wording. If verification is a condition precedent to coverage, the insurer may deny the claim outright. Some policies apply a higher retention or coinsurance instead, which is why that structure is worth negotiating before a loss. Have a licensed broker or coverage lawyer review the clause against your actual procedures.
Written by the Breachnodeops editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.