In this article, the FraudClassifier model in a press release issued by the U.S. Federal Reserve will be explored, along with its implications.
The Need for a FraudClassifier Model
The Federal Reserve System (hereinafter referred to as the “Fed”) classified a total of 12 types of fraud and provided scenarios and interpretations for each type.
Why did the Federal Reserve release its FraudClassifier model?
This FraudClassifier model is designed to cater to a wide array of stakeholders, including banks, payment gateways, law enforcement, and policymakers. It offers a comprehensive framework for categorizing payment fraud across various platforms and industries.
The intent is to provide a uniform method for identifying fraudulent activities. As online payments continue to experience exponential growth, the scope of fraudulent transactions is also expanding, with patterns becoming increasingly diverse. A simple Google search for ‘Types of Fraud’ yields over 20 distinct fraud categories, each further segmented based on specific business classification criteria.
When compiling statistics on incidents, the inconsistency in labeling for the same incident can result in significant expenses to standardize it. The Federal Reserve introduced FraudClassifier standards to mitigate these costs. Furthermore, based on the industry adoption roadmap for the FraudClassifier model, the objective is to harmonize the classification criteria across various sectors by 2024. Towards the conclusion of this article, we will assess the progress of this integration effort.
Existing Fraud Classifier Methods and How the Federal Reserve Classifies Fraud
To leverage the FraudClassifier model introduced by the Federal Reserve, it’s imperative to grasp its intricacies. This model categorizes fraud into 12 distinct types, relying on the responses to three fundamental questions: ‘Who’, ‘How’, and once again, ‘How’. The Federal Reserve’s FraudClassifier model guide map is as follows.
Existing fraud classifier models focus more on ‘how it occurred’ and determine fraud methods such as phishing, smishing, and identity theft and suggest measures to prevent them. This can be seen as the basic premise that Who = scammer.
On the other hand, the fraud model announced by the Federal Reserve distinguishes whether the payment was made by the principal (Authorized Party) or a third party (Unauthorized Party) . It was reported that, thanks to this, many small and medium-sized financial institutions were able to classify ‘fraud by the principal’ that had not been tracked before.
Scenarios According to the Federal Reserve FraudClassifier Model
Below are scenarios for each of the Federal Reserve’s FraudClassifier models.
- Authorized Party Was Manipulated – Products & Services Fraud
In addition to his recurring condo fee payments, Paul received an invoice for planned roofing repairs that appeared to be from his condo association. Paul sent a check to the address on the invoice. Soon after, he reached out to the condo association only to be told that the roof had just been replaced a couple of years ago and they had never sent an invoice.- Authorized Party Was Manipulated – Relationship & Trust Fraud
Joan fell in love with Fred via an online dating site. They decided to meet in person. A day before the meeting, Fred asked Joan to wire him $10,000 to get out of serious trouble. Joan wired the money to Fred. The next day, Fred did not show up for their date as planned. Joan made several attempts to reach out to Fred but never heard from him again.- Authorized Party Acted Fraudulently – Embezzlement
Tina, the company treasurer, could initiate ACH payments at her company. She decided to initiate and authorize a large ACH payment from her company’s account to an outside account under her control. She then withdrew those funds to settle her financial debts.- Authorized Party Acted Fraudulently – False Claim
Betsy used a computer at the town library to order new merchandise online, making the payment with an electronic check. Days after she received the goods, she called her bank and reported the purchases as fraudulent. Betsy stated that she did not place the order or make the payment, and suggested her account information or credentials must have been stolen and misused.- Authorized Party Acted Fraudulently – Synthetic ID
Fred opened a deposit account under a fabricated identity that included both stolen information as well as made up information. After a large deposit into that account, Fred initiated a wire transfer to an offshore account. He never transacted on the fabricated identity account again.- Unauthorized Party Modified Payment Information – Compromised Credentials
Steve set up an online recurring bill payment from his bank account while his roommate was nearby. His roommate later authenticated into Steve’s account using Steve’s ID/password and modified the recurring payment. Upon processing of the payment, funds were redirected to an account under the roommate’s control.- Unauthorized Party Modified Payment Information – Impersonated Authorized Party
Jim scheduled an online bill payment using his bank account information. A day later, Jake represented himself as Jim by successfully answering verification questions asked by the call center associate. Jake (as Jim) then instructed the call center associate to modify the scheduled payment, redirecting it to an account in his control. Upon receipt of the payment, Jake withdrew the funds.- Unauthorized Party Modified Payment Information – Physical Alteration
Susan hired a concierge service. Pauline showed up, indicating she worked for the concierge service. Susan asked her to buy her an item at the store, and gave her a signed check from her checkbook to cover the cost of the item. Pauline then changed the payee and amount information on the check, and deposited it into her account.- Unauthorized Party Took Over Account – Compromised Credentials
Using Greg’s login ID and password, Frank gained full access to Greg’s online bank account. Frank then proceeded to initiate several wire transfers from Greg’s account to a different account under his control. Once Frank drained Greg’s account, he changed the login password and logged out.- Unauthorized Party Took Over Account – Impersonated Authorized Party
Through a call center, Frank represented himself as Greg Jones using Greg’s stolen personally identifiable information. Frank then requested a large wire transfer to an account under his control and instructed the call center associate to close his (Greg’s) now empty account.- Unauthorized Party Misused Account Information/Payment Instrument – Digital Payment
Carl found an old check from Frank. Using the account information from the check, Carl initiated an online payment on the university’s website to cover the cost of his fall semester tuition bill.- Unauthorized Party Misused Account Information/Payment Instrument – Physical Forgery/Counterfeit
Dan, who works in the mailroom, stole blank corporate checks from the office. He wrote the checks to himself, forged the signature as an authorized representative of his company, and then cashed all of them.
Current Adoption Status of FraudClassifier Models
Is this FraudClassifier model being effectively utilized in the field? This model was announced in June 2020, with the goal of integrating it into various industries, including banks, payment operators, and e-commerce, by 2024. By now (2024), the model should be somewhat established, but it has not yet been fully integrated across all industries.
In the financial industry, the introduction of FraudClassifier models has positively impacted fraud detection, enabling the identification of previously untraceable frauds. However, in fields such as e-commerce, these models do not appear to be actively used yet. For example, Stripe, a major U.S. payment provider, categorizes payment fraud into six categories, including phishing, skimming, and identity theft. The company analyzes how fraud occurs and suggests preventive measures accordingly.
On the other hand, the Fed’s FraudClassifier model aims for a more fundamental and systematic defense by identifying the entities that initiate the fraud and how they behave. While granular fraud classifier models can be applied across various industries, including financial institutions and law enforcement, their integration into individual businesses can slow down implementation efforts.
Implementing FraudClassifier Models for Effective Detection
From the standpoint of individual businesses, the initial phase in embracing the Federal Reserve’s FraudClassifier model involves pinpointing the ‘who.’ Consequently, it’s essential to monitor customer logs throughout your business operations to swiftly identify any suspicious activities, such as unauthorized logins or alterations to account details.
However, for the majority of companies and organizations, attaining the requisite level of detection technology for practical implementation poses a significant challenge. Therefore, it’s crucial to judiciously utilize appropriate technologies or solutions in business operations. Criminal IP FDS is well-suited as a detection solution, offering monitoring capabilities for various activities, including payments, logins, and registrations. Its utilization of IP-based threat intelligence makes it particularly distinctive, as it can be seamlessly integrated across diverse industries.
Businesses can protect financial assets and customer data, elevate customer reputation, and strengthen compliance through fraud prevention measures. Criminal IP FDS offers detection outcomes like credential stuffing and risk pattern detection to accomplish this. Recently, governments and legislative bodies have also taken a proactive stance on fraud prevention. When fraud occurs, businesses may incur not only financial losses but also fines and penalties for regulatory non-compliance. Therefore, implementing effective technological solutions for preemptive measures is often a more cost-effective approach.
In relation to this, you can refer to the recommendations in the Neglected Basic Fraud Response Strategies: Insights from Credit Card Companies.
Source: Criminal IP(www.criminalip.io/)
Related Article(s):