Spoofed emails, real wires: the SEC put business email fraud inside payment controls

The SEC’s October 2018 report on nine companies that wired nearly $100 million to impostors placed the fraud in internal accounting controls. Its lasting value was the line it kept: controls over moving cash, not cybersecurity at large.

Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.

Two-minute summaryThe dek and the first sentence of each section, quoted from the article

The SEC’s October 2018 report on nine companies that wired nearly $100 million to impostors placed the fraud in internal accounting controls. Its lasting value was the line it kept: controls over moving cash, not cybersecurity at large.

  1. The week On Tuesday, October 16, 2018, the Securities and Exchange Commission published a report of investigation under Section 21(a) of the Securities Exchange Act of 1934.
  2. Authorization: approval rules that accepted an email as evidence Every issuer in the report had authorization rules.
  3. Master data: the vendor bank record as the real target The vendor schemes in the report were more sophisticated and harder to spot.
  4. Detection: finding the bad payment before the vendor does The report’s hardest facts concern detection.
  5. People: controls that depend on who is reading them The report’s last analytical point was that systems of internal accounting controls depend on the personnel who implement and follow them.
  6. Objections: guidance without charges, a theory the courts cut back, a problem for email security The first objection is that the report changed nothing.
  7. With hindsight The boundary held.

The week

On Tuesday, October 16, 2018, the Securities and Exchange Commission published a report of investigation under Section 21(a) of the Securities Exchange Act of 1934. Its subject was a set of cyber-related frauds against public companies and what they meant for the internal accounting controls those companies are required to keep. The Division of Enforcement had run the investigations in consultation with the Division of Corporation Finance and the Office of the Chief Accountant. The SEC’s release said the report coincided with National Cybersecurity Awareness Month.

The report described nine issuers that had each lost at least $1 million to what the FBI calls business email compromise. Two lost more than $30 million. Together the nine lost nearly $100 million, almost all of which was never recovered. The companies came from technology, machinery, real estate, energy, financial and consumer goods, and each had securities listed on a national securities exchange. The report cited FBI estimates that these schemes had caused over $5 billion in losses since 2013, with $675 million in adjusted losses in 2017 alone.

Two schemes were at work. In the first, impostors used spoofed email domains and addresses of a senior executive, such as the CEO, to instruct finance staff to work with a purported outside attorney and wire large sums to foreign accounts. The emails described urgent, secret deals, sometimes invoked government oversight, and went to midlevel staff who did not normally deal with the executive. In the second, criminals broke into the email accounts of the issuers’ foreign vendors, joined real conversations about purchase orders and invoices, and asked for changes to the vendors’ bank details, attaching doctored invoices. Procurement staff relayed the requests to the accounting staff who maintained vendor data, and payments on real invoices went to the criminals.

The losses were not discovered by the victims’ own controls in several cases. One company made 14 wire payments over several weeks, losing over $45 million, before a foreign bank raised an alert. Another paid eight invoices totaling $1.5 million over several months and learned of the fraud only when the real vendor complained about late payment.

The Commission decided not to bring enforcement actions against the nine. Instead it used the report to tell issuers that Section 13(b)(2)(B)(i) and (iii) of the Exchange Act, which require a system of internal accounting controls giving reasonable assurance that transactions and access to assets follow management’s authorization, apply to these threats. The issuers had procedures requiring payment authorization levels, management approval of outgoing wires and verification of vendor data changes. Staff read those procedures as allowing an email, on its own, to justify a large wire or a change to vendor banking data. One employee ignored a dual-authorization rule; another misread the authorization matrix as giving him authority reserved for the CFO. In two cases the recipients were chief accounting officers. Every issuer later sought to strengthen its payment authorization procedures and its verification of vendor data changes, and all of them enhanced training. The issuers that had learned of the fraud from outsiders also bolstered reconciliation and outgoing payment notifications.

The report closed by saying the Commission was not suggesting that every victim of such a scam had violated the law. Issuers, it said, must calibrate their internal accounting controls to the current risk environment.

The October 2018 report is remembered, when it is remembered at all, as the SEC saying cybersecurity belongs in internal controls. That reading is too broad, and the difference matters. The report did not say a company’s firewalls, passwords or email filtering were accounting controls. It said that the controls governing who may move company cash, and on what evidence, have to work when the evidence is an email that someone else wrote. In 2023 and 2024 the SEC pressed a broader reading, and a federal judge and two of its own commissioners pushed back along the line this report had respected.

This article argues that the report’s durable contribution was that boundary. It placed business email compromise inside the payment process, owned by finance, rather than inside the security program, owned by IT. For any organization that pays suppliers by wire or ACH, whether it is a listed bank holding company, a health system that issues municipal bonds or a public agency, the useful question is the one the SEC asked: where in the life of a payment did the control fail?

The report itself points to four places, and they divide the problem cleanly. The first is authorization: the rules for who may approve a payment and what evidence they must have. The second is master data: the vendor records, especially bank details, that tell the payment system where money goes. The third is detection: the reconciliations and notifications that reveal a bad payment after it leaves. The fourth is people: whether the staff who run the first three understand them well enough to apply them under pressure. Each has a different owner and a different failure. A final section takes the strongest objections.

A word on scope. Section 13(b)(2)(B) applies to issuers with securities registered under Section 12 of the Exchange Act or that file reports under Section 15(d). Credit unions, nonprofit health systems and government agencies are generally not such issuers. The frauds the report describes do not check that status, though, and the control design it implies applies to anyone who pays invoices. Where the law differs by type of organization, we say so; for the specific obligations of a given institution, organizations should confirm with their own counsel and auditors.

The statute itself explains why the four parts fit. Section 13(b)(2)(B) asks for reasonable assurance on four objectives, two of which the report relied on: that transactions are executed in accordance with management’s authorization, and that access to assets is permitted only in accordance with that authorization. The statute defines reasonable assurance as the degree of assurance that would satisfy prudent officials in the conduct of their own affairs. A wire sent on a forged instruction is a transaction executed without management’s authorization. Cash sent to a criminal’s account is access to an asset without it. The report did not need a new theory. It needed only to observe that forged instructions now arrive by email.

That framing also explains who should act. Controls over the authorization and execution of payments sit with the CFO, the controller and treasury. They are tested by internal audit and, for larger issuers, by the external auditor in the audit of internal control over financial reporting. Security teams contribute, but they do not own the payment process. The report’s nine cases show what happens when the payment process assumes that someone else has already verified the instruction.

The report also sat alongside a statement the Commission had issued in February 2018 on public company cybersecurity disclosures, which it quoted: cybersecurity risk management policies and procedures are key elements of enterprise-wide risk management. The February statement addressed what companies tell investors. The October report addressed what companies do with their money. Read together, they assigned disclosure of cyber risk to the disclosure process and protection of cash against cyber-enabled fraud to the accounting controls. That division of labor is the one the later cases tested.

Authorization: approval rules that accepted an email as evidence

Every issuer in the report had authorization rules. Payments required approval at set levels. Outgoing wires needed management sign-off. The frauds did not defeat those rules by bypassing them. They defeated them by supplying the evidence the rules asked for. An email that appeared to come from the CEO was treated as the CEO’s authorization. The report put it directly: the controls could be, and were, interpreted to mean that the compromised electronic communication was, standing alone, sufficient.

That is a design flaw, and it sits in the authorization rule, not in the email system. A rule that says “wires above a threshold require CEO approval” without saying how that approval is evidenced invites the email to stand in for it. The fix is to specify the channel. Approval for an unusual payment has to arrive through a channel the requester cannot forge from outside: a workflow in the payment or treasury system that requires the approver’s own authenticated login, or a call back to a number already on file, not one supplied in the request.

Unusual payments need their own rule

The fake-executive emails shared features the report listed: urgency, secrecy, a foreign transaction unusual for the company, minimal detail and a recipient who did not normally deal with the executive. Each of those is a property of the request, and each can be written into the authorization rule as a trigger. A payment to a new beneficiary in a new country, requested outside the purchasing system, by someone senior who asks for confidentiality, should route to a second approver who must confirm by an independent channel. The point of writing the triggers down is that the junior employee no longer has to judge whether to challenge the CEO. The rule challenges for them.

For a bank holding company, those triggers belong in the corporate treasury’s own procedures, separate from the bank’s customer-facing wire controls. For a health system, we would expect the riskiest flows to include capital project payments to contractors and payments to staffing agencies, where large invoices and changing payees are normal. For an agency, authorization runs through appropriation and certifying-officer rules that already demand evidence; the question is whether those rules say anything about how an instruction to change payee arrives.

Authorization matrices that staff can read

One issuer’s employee read the authorization matrix as giving him approval authority reserved for the CFO. That is a document-design failure. A matrix that requires interpretation will be interpreted, and under pressure it will be interpreted in the direction the requester wants. The test of an authorization matrix is whether two people reading it independently reach the same answer for the same payment. Where they do not, the matrix needs rewriting, or the payment system needs to enforce it so that no reading is required.

System enforcement is the stronger option. If the treasury platform will not release a wire above a level without approvals from named roles, the matrix is no longer a document to be read. It becomes a configuration to be tested. Auditors can test configurations. They cannot test whether a busy employee read a table correctly on a Friday afternoon.

Dual control that is actually dual

Another employee did not follow the company’s dual-authorization requirement and instead directed unqualified subordinates to sign off. A second signature is only a control if the second person has the authority, the information and the independence to refuse. A subordinate who signs because the supervisor asked adds a name, not a check. Organizations should look at who their second approvers are in practice for high-value payments, and whether any of them report to the first approver. Where they do, the pairing should change.

Evidence an examiner or auditor would expect

For each authorization control, an organization should be able to produce three things. The first is the written rule, including how approval is evidenced and which payments count as unusual. The second is the system configuration that enforces it, with a record of who can change that configuration. The third is a sample of released payments showing that the rule was followed, including the unusual ones. A control that exists only as a policy document, with no enforcement and no sample, will not satisfy a skeptical auditor, and in the 2018 cases it did not stop the money.

The same evidence serves organizations outside the SEC’s reach. A credit union’s supervisory committee, a health system’s audit committee and an agency’s inspector general can all ask for the rule, the configuration and the sample. The vocabulary differs. The test does not.

Thresholds that invite splitting

Fixed approval thresholds create a known target. A fraudster who learns that wires below a certain amount need only one approver can ask for several payments just under it. The report’s case of 14 wires over several weeks shows that repeated requests were not a barrier. Authorization rules should therefore look at cumulative payments to a new beneficiary over a period as well as at each payment alone. A second wire to an account first paid last week deserves the same review as the first.

Master data: the vendor bank record as the real target

The vendor schemes in the report were more sophisticated and harder to spot. The criminals had broken into the real vendors’ email, joined real conversations and attached invoices that looked right. Nothing about the request to change bank details was unusual except the change itself. The payments that followed were for real goods on real invoices. Only the destination was wrong.

This makes the vendor master file, and the process for changing it, the control point. A bank account change is the single edit that turns every future payment to that vendor into a payment to someone else. It deserves more scrutiny than any individual invoice, because its effect is multiplied across every invoice that follows.

Separating who asks from who changes

In the report, procurement staff received the requests and passed them to the accounting staff who maintained vendor data. That handoff is where a control can sit. The person who maintains the master file should not accept a change on the strength of a forwarded email, however legitimate the forwarder. They should confirm the change with the vendor through contact details already on file from before the request, ideally with someone other than the person who sent it. The report noted that every issuer afterwards sought to strengthen verification requirements for vendor information changes.

Where vendors maintain their own bank details in a supplier portal, the portal moves the problem rather than removing it. If an attacker controls the vendor’s email, they may also be able to reset the vendor’s portal password. Portal changes to bank details should trigger the same independent confirmation, and the account owner on the portal should be notified through a separate channel when banking information changes.

A cooling-off period and a first-payment check

Two further controls follow from the structure of the fraud. First, a newly changed bank account can be held for a set period before any payment is released to it, with the old and new details both visible to the approver. Second, the first payment to a changed account can be flagged for confirmation of receipt with the vendor. Neither control depends on spotting the fraudulent email. Both depend only on knowing that the account changed, which the master file always knows.

Who owns the vendor master file

Ownership of the vendor master file can be unclear, split between procurement, which onboards suppliers, accounts payable, which pays them, and IT, which runs the system. For this risk the owner should be a single finance role, with authority to refuse a change and a duty to report on changes. A useful monthly report lists every bank detail change, who requested it, how it was confirmed and by whom. In some organizations the number of changes in a month will be small enough for a controller to read the whole list. A change confirmed by replying to the same email thread that requested it is a failed control, whatever the policy says.

Onboarding is the first change

New vendor setup is the same control in a different place. Creating a vendor record with bank details is, in effect, a change from no account to an account. A criminal who cannot divert an existing vendor’s payments may instead try to create a plausible new vendor and invoice it. The confirmation step, the separation between requester and maintainer, and the first-payment check apply equally to new vendors. Procurement controls that require a purchase order before an invoice can be paid add a further barrier, since a fake vendor has no order to match.

Dormant vendors and stale contacts

The confirmation controls above depend on contact details already on file. Those details decay. A vendor’s finance contact leaves, a phone number changes, a company is acquired. If the contact on file is out of date, the person confirming a bank change may be pushed toward the new contact supplied in the request, which is exactly the channel the criminal controls. Periodic review of vendor contacts, done when nothing is being changed, keeps the confirmation channel trustworthy. Vendors that have not been paid for a long period can be deactivated, so that reactivating one, with new bank details, is treated as a new vendor setup rather than a routine payment.

For banks, the master data problem has a second dimension. A bank is also the receiving institution for other companies’ misdirected payments. Its own vendor master file needs the controls above. Its role as a receiver of fraudulent credits is a separate matter, governed by payment network rules and anti-money laundering programs, and should not be confused with its controls as a payer.

Detection: finding the bad payment before the vendor does

The report’s hardest facts concern detection. One company sent 14 wires over several weeks and learned of the fraud from a foreign bank. Another paid eight invoices over several months and learned of it from the real vendor’s complaint about nonpayment. The report observed that vendors may give customers months before treating a payment as late, which let the vendor schemes run.

A control environment that depends on the victim of the diversion, or on a bank abroad, to raise the alarm has no detective control for this risk. The report said the issuers that had learned of the fraud from third parties afterwards bolstered account reconciliation procedures and outgoing payment notification processes. Those two words point to concrete work.

Reconciliation that looks at destinations

Standard reconciliations confirm that payments recorded in the ledger match payments that left the bank. They do not ask whether the beneficiary account was the right one. A reconciliation designed for this risk compares the beneficiary account on each payment with the account history for that vendor, and lists any payment to an account first used recently. That list should be short enough in most periods for someone outside accounts payable to review.

Notifications to the people who would notice

Outgoing payment notifications send the vendor, at its contact of record, a message when a payment is released, showing the last digits of the destination account. The real vendor, seeing a payment it never received to an account it does not recognize, raises the alarm on the first payment instead of the eighth. For fake-executive schemes, a similar notification to the executive whose name was on the instruction serves the same purpose. The CEO who receives a note that a wire was released on his authority, for a deal he has never heard of, will say so.

Speed after detection

Detection matters because recovery depends on time. Once a fraudulent payment is found, the paying organization’s bank needs to know at once so it can attempt a recall and alert the receiving bank. The report noted that nearly all of the $100 million was never recovered. An incident procedure for payment fraud should name who calls the bank, at what hour, with what information, and should be tested. It belongs to treasury and finance. Security teams can help investigate the email side, but the money moves through finance’s relationships.

Measures that show the detective controls work

Detective controls can be measured. Useful measures include the interval between a payment’s release and its appearance on the beneficiary review list, the share of release notices that reach a valid vendor contact, and the time taken in exercises to reach the bank’s fraud desk after a suspected diversion. Each can be measured without a fraud having happened, and each tells the audit committee whether the organization would find the next one itself, or would again wait for a foreign bank or an unpaid vendor to tell it.

Two schemes, two detectors

The two schemes in the report leave different traces, and a detective design should cover both. The fake-executive scheme produces payments that are unusual in themselves: a new beneficiary, a foreign destination, no purchase order, an amount with no history. A review of payments outside the purchasing process catches them. The vendor scheme produces payments that look normal in every respect except the destination account. Only a control that compares the destination with the vendor’s history catches it. An organization that reviews only unusual payments will miss the vendor scheme, and one that reviews only bank detail changes may miss the executive scheme. The report’s case of eight invoices paid over several months shows how long the vendor scheme can run when no one compares destinations.

Using the bank’s services without outsourcing the control

Banks may offer corporate clients services that bear on these risks, such as payee verification, payment holds and alerts on unusual activity. They can help. But the report’s findings were about the issuers’ own controls, and a bank service is a tool inside those controls, not a replacement for them. An organization that relies on a bank alert should know what the alert covers, who receives it and what that person does next. For a bank holding company, which may be both the corporate client and the provider, the two roles should be documented separately so that the treasury function does not assume protections designed for customers.

People: controls that depend on who is reading them

The report’s last analytical point was that systems of internal accounting controls depend on the personnel who implement and follow them. The frauds succeeded at least in part, it said, because responsible staff did not understand the existing controls or did not recognize signs that the instructions were unreliable. Recipients asked no questions about transactions outside their domain, even when asked to make several payments over days or weeks.

Two details stand out. In two cases the recipients were chief accounting officers. And while most of the issuers had some training on controls and technology before the frauds, all of them enhanced it afterwards. Seniority did not protect anyone, and general training did not either.

Training for the role, not the workforce

The training that matters here is narrow. It is for the people who can release payments or change vendor data, and it should cover the specific scenarios the report described, using the organization’s own approval rules. A general phishing course teaches employees to distrust links. It does not teach an accounts payable clerk that a CEO request to wire funds abroad for a confidential acquisition must be refused until confirmed through the treasury workflow. The second lesson is the one that would have saved the nine issuers money.

We think organizations should measure this training the way they measure other controls: by testing. A periodic exercise in which a simulated executive request reaches finance staff, and the measure is whether the release procedure was followed, gives evidence an auditor can use. Click rates on phishing tests measure something else.

Permission to say no

The fake-executive emails relied on authority and secrecy. A midlevel employee asked by the apparent CEO to keep a deal confidential faces a real conflict. Organizations can remove the conflict in advance: senior leaders state in writing that no executive will ever ask finance staff to bypass payment procedures, and that staff who refuse such a request will be supported. Written in advance, that statement turns refusal from insubordination into compliance with policy.

Where staffing models raise the risk

Shared service centers and outsourced accounts payable change the picture. The staff applying the controls may be in another organization, under another manager, working through a ticket queue. The controls above still apply, but the contract has to carry them: who may change vendor data, how changes are confirmed, how payment releases are approved and how quickly suspected fraud is escalated. A health system or agency that has outsourced payables should check that its service provider’s procedures match its own authorization rules, and that the provider’s staff receive the role-specific training described above.

Executive assistants and shared mailboxes

Fake-executive schemes work best where finance staff are used to receiving instructions from or on behalf of executives. Assistants who manage executives’ email and calendars, and shared finance mailboxes that several people monitor, are part of the channel. Organizations should know which mailboxes can send messages on an executive’s behalf, and should tell finance staff that a payment instruction from an assistant or a shared mailbox carries no more authority than one from the executive, which under the authorization rule is none without the workflow approval.

Turnover and coverage

Controls that depend on knowledge are weakest during transitions. A new payables clerk, a temporary worker covering a vacation or an outsourced team in its first months may not know which requests are unusual. Role-specific training should be part of onboarding for anyone who receives payment release or vendor change rights, and those rights should not be granted until the training is complete. Removing rights when people change roles closes the other end.

What a useful test looks like

A simulated executive request is only useful if it resembles the real thing. It should reach the people who actually release payments, use the urgency and confidentiality the report described, and ask for an action the authorization rule forbids without workflow approval. The result to record is not whether the employee spotted a fake, but whether the payment procedure was followed regardless. An employee who suspects nothing but still routes the request into the workflow, where it stalls for lack of approval, has passed. An employee who spots the fake by luck but would have paid a better one has not. Testing the procedure rather than the instinct gives management and auditors evidence about the control, which is what the report asked issuers to maintain.

Objections: guidance without charges, a theory the courts cut back, a problem for email security

The first objection is that the report changed nothing. The SEC charged no one. A Section 21(a) report states the Commission’s views but does not create a new rule. Companies already had to keep internal accounting controls.

That is correct about legal force. The report’s effect was in where it placed the problem. Before it, a public company’s audit committee could treat business email compromise as a security incident to be reported by the CISO. After it, the same losses were evidence about the design and operation of controls over disbursements, which the audit committee, the CFO and the external auditor already own. The report told those people that the frauds were theirs.

The objection that the SEC later lost on this theory

The second objection is that the internal accounting controls theory did not survive. In 2024 a federal court dismissed the SEC’s internal accounting controls claim against SolarWinds, holding that the provision refers to financial accounting and cannot reach cybersecurity controls such as passwords and VPN settings.

That ruling cut back a broader theory, not this one. The court rejected the idea that controls over access to computer systems are internal accounting controls. It distinguished a case where the controls at issue were directly related to the integrity of financial transactions. Two SEC commissioners, dissenting from a different 2024 order, made the same distinction and cited the 2018 report as consistent with it: the 2018 frauds were about access to corporate cash through fraudulent transactions. The controls over who may release a wire and how vendor bank details change are controls over transactions. The boundary the 2018 report respected is the one the court drew.

The objection that email security is the real fix

A third objection holds that the frauds were a failure of email security, to be fixed with domain authentication, filtering and impersonation detection. Those measures help, and the spoofed domains in the report may well have been caught by them. They do not address the vendor schemes, where the email genuinely came from the vendor’s compromised account. And no email filter is perfect. The report’s logic is that the payment process should hold even when a convincing email gets through. Email security reduces how frequently the payment controls are tested. It does not replace them.

The objection that controls slow the business

A fourth objection is cost. Call-backs, holds on changed accounts and second approvals add time, and some payments are genuinely urgent. The report itself acknowledged that management must judge the steps and costs involved, quoting the 1977 Senate report on the same point. The answer is proportion. The controls above bite only on a narrow set of payments: those to new or changed accounts, those outside the normal purchasing process, and those that match the unusual-request triggers. Routine payments to established vendors through the purchasing system pass as before. An organization that finds the controls too slow for a legitimate urgent payment can create an expedited path with stronger, not weaker, confirmation.

The objection that small finance teams cannot segregate duties

A small credit union, a rural hospital or a county finance office may have only a few people in finance. Separating requester, maintainer and approver may seem impossible. The controls can be adapted. A board member or senior officer outside finance can serve as the second approver for unusual payments. The bank can be asked to require dual release for wires above a level. Vendor change confirmations can be logged and reviewed monthly by someone outside the team. These are imperfect, but they replace a control that depends on one person’s judgment with one that does not.

The objection that generated text and voices outrun these controls

A final objection is that the 2018 frauds were crude, with spelling errors the report noted, and that requests written or spoken by AI tools now make red flags disappear. The FBI’s 2025 report recorded business email compromise losses involving AI. That trend weakens controls that depend on spotting a bad email. It does not weaken the controls this article describes, because they never relied on the quality of the request. A call-back to a number on file, a workflow approval behind the approver’s own login, a hold on a changed account and a destination comparison all work the same way whether the request is clumsy or flawless. If anything, better forgeries strengthen the case for moving the control out of the inbox and into the payment process, which is where the 2018 report placed it.

What we would tell a CFO today

Assign each of the four control points a named owner in finance: authorization rules and their enforcement in the payment system, vendor master data changes, beneficiary-focused reconciliation and notifications, and role-specific training and testing. Ask internal audit to test each one against the two scenarios in the 2018 report. Give each owner a short set of measures to report quarterly: bank detail changes and how each was confirmed, payments to new or changed accounts and how each was approved, the result of the latest simulated executive request, and the time taken to reach the bank in the latest recall exercise. Then ask one plain question of every payment process: if an email that looked exactly right arrived tomorrow, what in this process, other than someone’s judgment, would stop the money?

With hindsight

The boundary held. On July 26, 2023, the SEC adopted rules requiring registrants to disclose material cybersecurity incidents on Form 8-K, generally within four business days of determining materiality, and to describe their cybersecurity risk management and board oversight annually. Those were disclosure rules. They left the internal accounting controls provision where it was.

The SEC then tested a wider reading. In a settled order of June 18, 2024 against R.R. Donnelley & Sons, the Commission found a violation of Section 13(b)(2)(B)(iii), treating the company’s information technology systems and networks as the assets whose access controls were deficient. The order, as the dissent described it, faulted the company’s procedures for reviewing alerts from its intrusion detection system and its managed security provider. Commissioners Peirce and Uyeda dissented. They wrote that the provision covers assets that are the subject of corporate transactions, and that the 2018 report was consistent with that distinction because its schemes involved access to corporate cash through fraudulent transactions.

On July 18, 2024, the U.S. District Court for the Southern District of New York dismissed the SEC’s internal accounting controls claim against SolarWinds. The court held that “system of internal accounting controls” refers to financial accounting and cannot reasonably be read to cover cybersecurity controls such as password and VPN protocols. It noted that the provision was enacted in 1977 as part of the Foreign Corrupt Practices Act, long before cybersecurity was a business concern. On November 20, 2025, the SEC filed a joint stipulation with SolarWinds and its chief information security officer to dismiss the rest of the case with prejudice, saying the decision was in the exercise of its discretion and did not necessarily reflect its position on any other case.

The fraud did not go away. The FBI’s 2025 Internet Crime Report counted 24,768 business email compromise complaints with reported losses of $3,046,598,558, against $2,770,151,146 in 2024 and $2,946,830,270 in 2023. It said businesses reported losses over $30 million in 2025 to business email compromise involving AI. Its Financial Fraud Kill Chain process, which tries to freeze funds after fraudulent transfers, handled 3,900 incidents in 2025 and froze $679,013,183.

Payment networks moved toward the 2018 report’s detective controls. Nacha’s ACH rules now require risk-based processes reasonably intended to identify entries authorized under “False Pretenses,” a term Nacha says covers business email compromise and vendor impersonation. The first phase took effect on March 20, 2026 for all originating banks and for originators and third parties whose 2023 ACH origination or transmission volume exceeded 6 million entries; the second, effective June 19, 2026, applies to all other non-consumer originators and third parties. Covered parties must review those processes at least annually. Nacha says the rules do not prescribe specific methods, and that a risk-based approach should not be used to conclude that no monitoring is necessary.

What held up was the report’s placement of the problem in the payment process and its focus on authorization, vendor data, reconciliation and training. What did not hold up was the hope, visible in the SEC’s later cases, that the same provision could carry a general cybersecurity mandate. As checked on November 14, 2026, it applies to the controls over transactions, which is where the 2018 report put it.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • business email compromise
  • internal accounting controls
  • SEC
  • payment fraud
  • vendor master data