Thwarting Fraud Before It Happens
by Ann McDonald
Banks are turning to identity scores to determine if people are who they say they are. How is it working?
Move over credit scores – there’s a new type of score on the block: identity scores that can calculate the risk of fraud before a consumer opens an account. Vendors say their use is bound to grow because of mounting concern about identity theft and fraud. Last year, for the fifth year in a row, identity theft topped the Federal Trade Commission’s list of consumer complaints, accounting for 39% of all complaints filed with the agency in 2004.
Several companies over the past three years have introduced identity or fraud scores – numerical scores that look at many different variables to verify an individual’s identity and determine fraud risk. Although some industry analysts say the technology still has not drawn a large following, vendors say that the tools are beginning to pick up steam as financial institutions and credit grantors look to reduce their fraud losses.
Banks and lenders have used a variety of methods in the past to verify a customer’s identity, including manual credit checks or using external data services to check identities, but with identity theft on the rise the importance of interactive services to stop fraud at the point of application is increasing. The problem with old techniques for verifying identities, according to Mike Cook, co-founder and vice president of ID Analytics, is that they take too much time and businesses can lose good customers because of bad or incomplete information in the databases they search.
San Diego-based ID Analytics introduced its ID Score about two and half years ago. The score is generated from the company’s ID Network, a fraud-detection system that has application information from more than 1 billion transactions, according to Cook. Based on that network of identity attributes, such as Social Security numbers as well as names and addresses, the company generates a risk score to measure the likelihood that someone is who he or she claims to be when opening an account.
The score also identifies synthetic identity fraud, which occurs when an identity contains some valid data elements, along with invalid ones. This type of fraud often goes undetected because the fraudster establishes an apparently authentic identity and follows patterns of behavior that appear legitimate – such as making monthly minimum payments on a credit card bill.
Customers can send applications to ID Analytics in batches or access the service interactively immediately after an application has been processed. After scouring the ID Network, ID Analytics returns a score from 0 to 999. If the score is higher than the creditor’s risk level, the creditor is notified and can contact the consumer to try to verify his or her identity. The score is powered by the company’s Graph Theoretic Anomaly Detection system – a pattern recognition technology that discriminates between legitimate and suspicious identity behavior.
Spreading the Word
Working to make the ID Score the industry standard for risk assessment, Cook says, ID Analytics has entered into resale partnerships with several prominent companies – including Visa, Equifax, Intersections Inc. and Carreker, a technology provider for financial institutions. ID Analytics customers, Cook says, include five of the top ten credit card issuers and two of the top four retail lenders.
Susan Hall, vice president of risk products for Visa, says that a joint study with ID Analytics projected that using the ID score in combination with Visa’s clearinghouse service – which flags high-risk applications based on information from the three major credit bureaus and public records – would help member banks identify 17% to 34% more fraudulent applications.
Visa is in the process of launching just such a customized version, ID Score Plus – a combination of the ID Score and Visa’s clearinghouse service – for its member banks. Pricing for the product has not been announced, but Visa is in talks with many banks about the service, Hall says. “Our members are very much aware of the technology,” she says. “The real advantage here is the fact that we are able to deliver access [to the ID Score] through our channel, requiring little development on our members’ part.”
The strength of the ID Score, says Avivah Litan, vice president and research director for Stamford, Conn.-based Gartner Research, is that it makes decisions about fraud based on a cross-industry shared database that can catch fraud across many different sectors, from telecommunications to banks and retailers. “The best way to catch fraud is to look at the patterns of transactions of that group of users and to do that at the identity level [when opening an account],” she says.
The only downfall Litan sees with the ID Score is that its database is more focused on information from new account openings. It would be a more powerful tool, she says, if it had information on consumers’ transaction behavior after accounts are opened.
Fair Isaac also offers an identity scoring product, Falcon ID Score, which aims to catch fraud at every point of customer interaction. The technology allows a lender to review only 10% of applications to find the most fraud, says Ted Crooks, vice president of global fraud solutions for the Minneapolis-based company. “The trick is to find as much fraud as you can in as few cases as possible,” he explains. “Think of the system as trying to explain why a customer is good, and if it fails, it gives a high score.”
The first step for the system is to detect fraud by comparing credit bureau information to application information and return a score of 0-1000. If a high score is returned for the application, the company will do a manual verification, asking a series of questions to verify the identity of an individual. As part of Falcon ID, Fair Isaac provides customers with out-of-the-wallet questions that draw from a variety of data sources, not just credit reports, to make the questions more difficult for a fraudster to answer.
“We may check and see if we can find them in a boat owners’ directory and ask them about their boat,” Crooks says. “We use sources that don’t apply to everyone and ask a lot of questions a fraudster wouldn’t expect.”
The system draws from many non-traditional data sources, such as genealogy records or geographical information on that customer. Data providers TARGUSinfo and Qsent have also recently signed on to provide information for Falcon ID.
Crooks declines to reveal how many customers use the Falcon ID product, but he does say that the company has customers in many different industries, including telecommunications companies and financial institutions. Fair Isaac recently took steps to expand its client base through an arrangement with MicroBilt to resell the score to its customers, including many small to mid-sized lenders.
Ken Hill, president of MicroBilt, says his firm looked at several types of fraud prevention technologies before deciding to resell the Falcon ID product. The company was impressed with the combination of identity scoring and the question-and-answer component that draws on data from sources other than credit reports.
MicroBilt is working on integrating the product into its system and hopes to have some customers using the technology by the fourth quarter of this year. Hill says MicroBilt is working at keeping the yet-to-be-determined price for the service competitive with other offerings on the market. Both ID Analytics and Fair Isaac declined to discuss how they price their offerings.
“I think once customers understand the product and how it will help them, they will be interested,” Hill says. “The main benefit is that it streamlines the identity verification process. It helps to ensure that the person you are doing business with is the person they say they are, so you are not exposing yourself to fraud.”
Experian is another firm engaged in fraud scoring. Many of Experian’s customers say they have difficulty keeping up with the constantly changing nature of fraud, says Heather Vernon, director of product development for the Costa Mesa, Calif.-based company. The company offers a Fraud Shield Score, which uses information from credit reports to determine the risk of fraud at the time of application and provides a combined credit and fraud risk score.
This month the company is launching Precise ID, which provides a single point of access for the firm’s data to determine fraud on the front end. The product aims to prevent first-party fraud, which occurs when an individual opens an account in his or her real name and then runs up huge amounts of charges with no intention of repaying them, as well as third-party fraud, which occurs when someone opens an account in another individual’s name.
Experian uses many different sources of data, including white page information, public record data and its own marketing database to verify identities. The technology also scours 450,000 known fraud records in the National Fraud Database, a shared database of information from creditors and financial institutions. The information is incorporated into a model that produces a score indicating the risk level for fraud. If a high score is returned, Precise ID can also indicate what type of fraud it may be, whether it follows the pattern of a synthetic ID fraud, the actions of a fraud ring or an actual ID theft.
A few of Experian’s clients are already implementing the technology, Vernon says, and many more are slated for next year. Customers like having a single point of integration for all of Experian’s data, as well as a platform that keeps up to date on the constantly changing nature of fraud, she says.
Hurdles Remain
Although the technology is catching on and more vendors are developing fraud-scoring tools, it may still take a while before fraud scoring is an industry-wide practice, Gartner’s Litan says. Fair Isaac and ID Analytics declined to give figures on how many users they have or to provide the names of specific customers, saying that users do not like to discuss their use of fraud detection tools for fear of tipping off fraudsters.
Litan says that relatively low fraud losses and concerns about sharing information have kept some financial institutions away from the technology. “Financial institutions definitely want this technology, but they don’t necessarily want it as a shared service,” she says. “A lot of financial institutions are worried about privacy issues when sending data to a third party. Also, for competitive reasons they may be hesitant to share all their new prospects with a third party.”
If collections and credit risk departments communicated better and realized that many accounts being sent to collections are identity theft cases, she says, it might hit home how important it is to identify fraud at the point of application. Until then, fraud-scoring vendors will have to work to get the word out to more lenders about what the technology can do for them.
It's time to protect your rights with an experienced attorney, not a paralegal.
