News Release

New way for hackers to target bank accounts discovered by UMass Amherst researchers

During a bank-to-bank transfer, money can be deposited into one bank account without ever being withdrawn from the other account

Reports and Proceedings

University of Massachusetts Amherst

Taqi Raza

image: 

Taqi Raza, Assistant Professor, Electrical and Computer Engineering at UMass Amherst

view more 

Credit: UMass Amherst

AMHERST, Mass. — Researchers at the University of Massachusetts Amherst have discovered how, during a bank-to-bank transfer, money can be deposited into one bank account without the funds ever leaving the second bank. This new class of cyberthreat is called Credit-Without-Debit (CWD) attacks and the researchers have demonstrated three ways hackers can use it to steal money.

 

“Our goal is to strengthen financial security,” says Taqi Raza, assistant professor in the Riccio College of Engineering at UMass Amherst. “The way we strengthen it is we find the loopholes and tell the banks, ‘okay, fix this problem.’”

 

Unlike previous research by this group, which focused on credit card and digital wallet vulnerabilities, this new study examines debit transactions, or ACH transactions. This system permanently moves money between banks—as opposed to creating a debt—and enables any bank-to-bank transaction, including direct deposit of salaries, bill payments (including credit card bills), mortgage services, and withdrawals from digital wallets and various payment apps.

 

The way an ACH transaction should work is that transfers are processed through multi-stage interbank clearing and settlement. Consider the situation where a person has bank accounts at two different banks and wants to move money from Bank A into Bank B. The person initiates the transfer from Bank B, which pulls the funds from the account at Bank A.

 

The transaction is authorized when Bank B recognizes that Bank A has a valid routing and account number and then authorizes the transaction. There is a two-day payment processing window for Bank A to transfer the money, so to expedite access to funds, some banks provisionally post funds in their customer’s account (in this example, Bank B posts these funds). Bank A checks that its customer’s account has sufficient funds, that there is valid authorization, and that there are no account holds, and then sends the money to Bank B. The interbank settlement is complete.

 

However, in a CWD attack, the attacker finds a victim with a frozen account, meaning the account can receive money but not send it. They link the victim’s account to their own and then pull the victim’s funds into their own account. The attacker’s bank posts the provisional funds. However, because the victim’s bank is frozen, the settlement payment is never made. In the two days it takes for this payment to be processed and rejected, the attacker has withdrawn the provisional funds, resulting in a direct financial loss to the bank. The researchers have created an interactive graphic to demonstrate how this attack works.

 

“The key root cause of all these attacks is that banks have treated distinct security checkpoints as interchangeable,” explains Raza.

 

Once they discovered the vulnerabilities, the researchers coordinated with affected banks to resolve the issues, which have not been reproducible since. The team presented their paper on the work today at the ACM Conference on Computer and Communications Security (CCS), 2026.

 

The researchers found two additional instances where this credit-without-debit logic resulted in successful attacks. One attack used microdeposits from third-party payment apps used to verify bank account ownership; the other paid off a credit card bill using a frozen account.

 

“We focused on what the user can observe and what they can influence from their banking apps,” says Ali Malik, Ph.D. candidate in electrical and computer science at UMass Amherst. “They can initiate payments, they can cancel those at random moments, they can observe the progress of that transaction as it moves from processing, to settle, to cancel, to rejected. It’s simple stuff. We defined our threat model around that.”

 

Even with these vulnerabilities discovered, Raza recommends that consumers always follow security best practices, such as enabling two-factor authentication, monitoring bank statements for fraudulent activity and turning on text, email, or banking app notifications whenever a debit or credit happens on an account.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.