Expert Insights

Real-time debit fraud monitoring

By Kevin Christensen

With every fraudulent swipe of a debit card, a financial institution loses money. The risk only increases as debit fraud rises and the criminals who perpetrate payments crimes become smarter.

Community banks have two primary software systems that can effectively protect against debit card fraud in either real-time or near real-time.

The first debit card fraud monitoring option involves the use of a near real-time neural network scoring engine. Essentially, this system gives every debit transaction a risk score depending upon its riskiness, based on various factors determined by the cardholders’ financial institution. The higher the risk, the higher the transaction’s risk score. If a score is high enough, a near real-time system would block the card. The initial (and potentially fraudulent) transaction would still be approved. However, such security software would then decline any subsequent transactions.

Real-time transaction-risk scoring engines also exist that would approve or decline each and every transaction based on a transaction risk score. Some of these software systems allow financial institutions to select community banks should consider processing multiple clearing windows throughout the day, rather than processing just one batch of transactions at the end of the day.

Meanwhile, while knowing your customer is important, knowing the capabilities of your IT vendors is also paramount. Like they do for other IT services, community banks providing mobile banking should perform careful due diligence on all their mobile vendors, particularly the vendor writing the bank’s mobile banking gateway app, Whaley says. How is the app being coded? How trustworthy is that software developer? How should the bank communicate to the public which transactions they want to monitor in real-time. Typically, card-not-present transactions, transactions over $100 and international transactions have a much higher fraud risk and should always be scored in real-time. However, community banks should also review whether it makes business sense for them to score all their customers’ debit transactions in real-time.

If a community bank uses a real or near real-time system, a fraud detection specialist would contact the cardholder to review recent activity and reissue the card (if the initial transaction was fraudulent) or unblock the card if the transactions were valid.

In emergency situations, financial institutions can also react with card-blocking technology in real-time, allowing them to shut down entire groups of transactions with the click of a button. If, for instance, a community bank receives a series of fraud reports originating out of grocery stores in central Illinois, it could block transactions at all grocery stores in the central part of the state (based on ZIP codes). Thus, all grocery store transactions out of central Illinois that its app is trustworthy?

Whaley points out that mobile payment services also heighten the need to monitor customer activity across channels to guard against electronic forms of traditional check kiting. To help minimize potential duplicate deposit problems, some mobile deposit vendors are also working with banks and check cashers to track and share when items have been processed, Krebs says. Such collaboration would shorten the potential damage and time frame that a fraudster could use to “beat the clock” and manipulate a transaction posting process.

“If a bank is going to offer mobile deposit capture, I would recommend will be blocked at the point of sale. This “shotgun” approach can save banks an untold amount of money. However, valid transactions will also be declined for all cardholders buying groceries in the ZIP code. Community banks should take a close look at their overall fraud losses to help craft their most effective fraud scoring strategy. If losses are historically concentrated in one segment (for example, e-commerce transactions) then a conditional real-time scoring strategy may provide the best return. However, if there is no discernible pattern of fraud losses, then scoring 100 percent of transactions in real-time will likely be most effective.

As always, criminals continue to modify their attacks to exploit vulnerabilities in the payments system. For that reason, community banks should monitor and review the effectiveness of their debit transaction fraud methodology. With the appropriate strategy, neural network scoring and card-blocking systems, whether operating in real-time or near real time, can still be very effective at fighting debit card fraud. a partnership, as knowledge sharing is really going to be the key for preventing this kind of behavior,” Pascual adds.

For the Bank of America fraud case in Kentucky, the first money order cashed and cleared through Western Union was the one to be paid, Whaley says. If the money order that was deposited through Bank of America cleared at Western Union before the money order was cashed at Kroger, then Kroger would have taken the loss, he points out. “It’s really a race to get the money.”

Katie Kuehner-Hebert is a financial writer in California.