Operationalizing Real-Time Fraud Detection
1 like517 views
This document discusses operationalizing real-time fraud detection through an integrated decisioning platform. It summarizes fraud loss trends showing check fraud exceeding $1 billion for the first time in 2008. Regulations now allow next-day check availability, benefiting customers but increasing deposit fraud as funds can be withdrawn quickly. The document advocates a unified fraud prevention approach using image analysis and decisioning platforms to enable real-time or near real-time fraud detection across check processing, deposits, and other transactions. This represents a shift from traditional day-two fraud detection practices hampered by paper-based processing.
Operationalizing Real-Time Fraud Detection
- 1. Fraud Prevention Operationalizing Real-Time Fraud Detection White Paper Integrated decisioning platform provides migration to day-one check fraud prevention Co-authored by Orbograph and The Santa Fe Group February 12, 2011
- 2. Introduction Continuous waves of check fraud attacks have heightened the sensitivity of financial institution (FI) executives. The growth of these crimes have forced FI’s to take defensive measures to minimize losses and protect their customers. Unfortunately, FI’s are hampered by limited capital budgets and staff reductions. Because FI resources are limited, gaps persist, allowing fraudsters to penetrate older systems built on outdated methodologies. The result is unwanted fraud risk. Are we in effect perpetuating check fraud by not moving fast enough to new technologies and solutions? The ABA 2009 Deposit Account Given the now common use of images and check truncation, it Fraud Survey is surprising that so many banks still view fraud detection showed check through the lenses of traditional check or siloed processing fraud losses practices with disparate case management and decisioning exceeded $1 billion for the first time. platforms for on-us, deposit, and ACH fraud prevention. This white paper offers an offensive game plan for financial institutions to solidify their defenses, minimize losses, maximize efficiency, and build for the future with enterprise-wide prevention strategies. I. Fraud Loss Trends and the Persistence of Paper The American Bankers Association reported in its 2009 Deposit Account Fraud Survey that, in 2008, check fraud losses exceeded $1 billion for the first time. Responding financial institution fraud managers reported that the top three loss categories were: returned deposited items, also known as deposit fraud (35%); counterfeit checks (30%); and forged makers’ signatures and endorsements (22%). This was reinforced by data from the Association for Financial Professionals’ 2010 Payments Fraud and Control Survey showing the primary sources of loss among commercial bank customers were: counterfeit checks using the organization’s MICR line data (72%); alteration of payee names on checks issued by the organization (58%); and alteration of dollar amount on checks issued (35%). 1 See http://www.aba.com/Surveys+and+Statistics/2009_Deposit_htm 2 See http://www.afponline.org/pub/pdf/2010_Payments_Fraud_Survey.pdf
- 3. In the ABA survey, more than half of respondents experienced ACH originator fraud during 2008 in frauds that typically targeted individual and small business accounts. As ACH volumes flatten and net ACH fraud dollars remain significantly less than on-us and deposit check fraud losses, it’s clear ACH volumes and losses will undoubtedly continue to grow, however it will be a long time before they approach the current and expected check and loss volumes. By comparison, it becomes obvious that investment in checks still promises good return. A 2010 customer survey by Orbograph asked financial institutions how their check fraud losses compare to the ABA median. Interestingly, 66.5% indicated that fraud losses were less than the ABA median. However, 84% out of 256 A 2010 Orbograph respondents believed an image-based fraud prevention customer survey strategy would provide moderate to strong value. showed 84% out of 256 respondents In addition to these eye-opening industry numbers, other risk believed an image- considerations for the paper-channel should be evaluated: based fraud detection strategy would Large-scale adoption of remote deposit capture will provide moderate to allow hundreds of thousands of businesses to feed strong value. images into the check clearing system virtually unchecked. Least cost routing options create opportunities for additional cross channel fraud when checks convert to ACH and move to Regulation E (90-day return rules). These extended windows make it more difficult to track fraud as well. Mobile capture for depositing checks provides great convenience, yet adds risk for potential exploitation by criminals. Adoption is predicted to climb in 2011. Effective fraud managers must be able to identify, interpret and prioritize these macro trends but account for micro level statistical variances including, urban vs. rural, organized crime presence and an ever-changing economic environment. However, analysis is not enough. Continual improvements using best practices and new technologies can enable FI’s to drive prevention/detection rates significantly higher.
- 4. II. A Tectonic Shift…In Thinking Before the enactment of the Check 21 Act in 2005, check fraud detection was a “day-two” function by necessity. There was no time to look for fraud during nightly check processing. Instead, bank check processing departments were consumed by two objectives: 1) to process all outgoing paper checks (those drawn on other banks) in time to meet the exchange deadlines each night; and 2) to process all the on-us paper checks (those accepted in-house the previous day and those received through in-clearings from other banks at night) in time to meet posting deadlines each morning. Banks could not afford to The banks’ “mission: critical” was to ensure that customer jeopardize the accounts were up to date before opening for business in the completion of morning. Each evening millions of checks were transported this intensive between hundreds of exchange points and thousands of banks paper-movement process for a to achieve the efficient nationwide clearing of checks upon handful of checks which the US economy has depended for decades. Banks that might not be could not afford to jeopardize the completion of this intensive legitimate. paper-movement process for a handful of checks that might not be legitimate; there was too much float at stake. Fraud managers finally had their opportunity when the on-us paper from the day before had been processed and posted. At that point, the race was on to locate fraudulent checks before midnight of that second day, the last clear chance for a bank to return an item to the depositing bank without being responsible for the loss. On the “returned check” side of the house, banks were again impeded by the return of the paper checks — checks they had previously accepted that were drawn on other banks. While in theory returning a check should take no more than two days, the complexity of the paper process created a lot of opportunity for checks to fall through the cracks, leaving the return of checks to wait until the third or fourth day after they had been accepted. This created an opportunity for criminals, who could
- 5. move or acquire the cash, sticking the accepting bank with the loss. Even in 2009, this concern continued. (See appendix A) Fast forward to today. Nearly all checks are imaged early in the banking process. Many are truncated right at the branch with increasing numbers imaged and entered into the clearing process at the customer’s office through RDC. The need to move paper through the clearing process to effect payment is a thing of the past. Nearly all of that paper transfer is accomplished electronically in a fraction of the time. Regulatory changes and technology developments are redefining fraud risk mitigation strategies as well. Regulation CC changes in 2010 provided next-day check availability and although positive from many operational and customer vantage points, fraudsters can work comfortably in many markets within a two day window by withdrawing funds next day via branch or ATM. The result is an increase in deposit fraud. Why? Regulation CC swings the pendulum — and the advantage — back to the paying bank. Institutions with strong on-us check fraud technology simply return potential counterfeits, irregular signatures (forgeries) and alterations to the BoFD (Bank of First Deposit) within the required processing window. By the time the BoFD receives the return, the money is gone. The paying bank has protected itself while the BoFD takes the loss. “This change is a tectonic shift in check processing and is not Jodi Pratt only related to the speed of processing,” says Jodi Pratt of Jodi Senior Consultant with Pratt and Associates & The Santa Fe Group. The Santa Fe Group III. A Unified Approach We now have the tools, the capacity, and the capability to move check fraud prevention into the mainstream of transaction processing and integrate it into the check and payment processing function so that the two can occur simultaneously. Platforms are now available to manage return items, exceptions, and fraud suspects from the same user interface. Additionally, multiple payments can be added, linking account activity and address cross channel fraud. Couple a platform strategy with additional data elements available through check
- 6. images, and detection may finally develop into fraud prevention. The overwhelming majority of banks have yet to seize this opportunity. The limitations that pushed fraud detection out of day-one processing no longer exist, yet banks continue to behave as if day-one detection is still out of reach. Most fraud managers still operate under the now-unnecessary constraints of a batched, day-two world. This is inefficient, certainly, and it allows criminals who’ve successfully exploited paper processing in the past to continue their work. What if banks began using today’s expedited check processing to transform check fraud detection and actually achieve the Holy Grail of day-one — and in some cases, day-zero — fraud detection? Can an FI overcome old habits and reap the full fraud-prevention benefits of imaging and new platforms? “With several simple steps, FI’s can step into the next generation: operationalizing real-time or at least “near real- time” fraud detection and prevention,” summarized Bob Jones, of RW Jones and Associates. IV. Technology and Operational Considerations Robert W. Jones Senior Consultant with The Santa Fe Group A well-rounded fraud-prevention program that uses all the available data points hasn’t always been easy to achieve. Initially, data points were generated from processing data and MICR line information feeding data analytics with basic statistical analysis. The result: lower than desired detection rates and high-false-positive rates. Today, decision data is generated by image analysis algorithms applied to imaged signatures, check stock formats, and field level characteristics. The key is a unified approach, leveraging analytics and image analysis on an integrated decisioning platform that is integrated into both the inclearings and over-the-counter (OTC) teller or branch capture environments to enhance pre-post fraud filtering.
- 7. Inclearings: Most checks cleared through the Federal Reserve and industry exchange networks are in image cash letters. Images arrive several times throughout the day depending on rules of each network. Running cash letters through an integrated check processing fraud detection system before posting offers an early warning of suspect items. Fraud is potentially caught earlier in the process, preventing on-us losses that previously were not discovered until it was too late to return items to cover a loss or stop subsequent intraday withdrawals. Adjusting in-clearing processes this way can transform check fraud management by making fraud prevention a mainstream function (as opposed to a post-clearing or auxiliary function). Inclearing processing offers other benefits too, including enhanced customer service by intercepting suspect items before they post to customers’ accounts and preventing legitimate items from being rejected. Payee name verification (PNV) is also a value-added feature reduce fraud. PNV can now be run Day 1 or Day 2 as an additional prevention method using customer generated payee issue files. Joe Gregory Vice President Marketing Over-the-counter (on-us): Transaction fraud filters, Orbograph based on the limited numerical data from the MICR line can offer the initial baseline of information about account behavior: the volume of checks written, the value of those checks, their serial numbers, and their timing. This data alone, however, is inadequate to make high-quality, low-false-positive suspect decisions, keeping many banks from implementing it at the teller line. Today, images captured at the back counter of a branch, or at the teller window, can be used to improve fraud identification rates, allowing for real-time fraud detection that keeps cash from crossing the teller account on cashed checks and split deposits. “These elements, combined with an intelligent decisioning engine, can now score an individual transaction within two to three
- 8. seconds,” stated Joe Gregory, Vice President of Marketing at Orbograph. Transit: Deposit fraud is generated from transit items. Solutions in this area are still evolving as many FI’s are now using “black list” servicers who provide access to fraudulent account lists. Although improving, institutions are not leveraging image-based data elements for fraud prevention. Several strategies can be deployed to help this area. FI’s can now create transit account profiles for matching payable transit items against during day one processing. This approach leverages an image archive which is now underutilized and allows FI’s to put temporary holds on funds. Profiles can be created from FDIC alerts for official checks which are involved with active fraud schemes. Image analysis can also reduce risk by identifying remotely created checks (RCC) and matching payee information to active “black lists” currently “Because one FI’s effecting the FI. Lastly, analytics on deposit trends at the “on-us” losses are another’s “deposit account level are the last line of defense and are fraud” losses, FI’s improved when merged with these advanced image need to understand techniques. the benefits of reciprocity,” stated “Because one FI’s “on-us” losses are another’s “deposit Jones. fraud” loss (and vice versa), FI’s need to understand the benefits of reciprocity,” stated Jones. There is substantial value to be gained by both parties in receiving reliable return information on the same day or early the next day. By cooperating with a common exchange partner, FI’s can reduce deposit fraud by the paying bank returning a notification of doubtful collectability to the BoFD. By taking this approach, an improved fraud network could spread quickly across the country, strengthening the entire banking community. 3 Other instruments: Once banks have invested in real- time on-us check fraud detection, it’s relatively easy to extend that functionality to additional areas of fraud loss. While losses associated with savings withdrawals, 3 For evidence of this, see other successful fraud collaborations that have occurred over the last 15 years, such as the BITS Fraud Steering Committee, the American Bankers Association Statistics Committee.
- 9. convenience checks on lines of credit, and cash advances have not been adequate to justify a special fraud detection solution, the infrastructure is in place to accept and analyze signatures on these documents. The incremental cost of including verification may now be justifiable as checks become harder for crooks to negotiate. V. Conclusion The 2010 Regulation CC changes on next-day check availability demand that banks focus on superior on-us fraud prevention. This is best accomplished when the foundation for case management is based on a common decisioning platform. This platform becomes the decision layer for a tightly integrated transaction and image-based fraud analysis engine. When “best of breed” performance feeds a common platform, the primary objectives of detection and efficiency can be achieved. Moving into the future, the decision platform incorporates additional payment channels under this umbrella and bridges the gap to day one fraud by supporting same day interbank fraud prevention.
- 10. APPENDIX A: ABA Deposit Fraud Report - 2009 Edition Average Time to Return Check to BoFD