In my last post about synthetic fraud, we looked at why many organizations are starting to pay attention to, plan for, and—most importantly—safeguard against rising levels of synthetic fraud. In this post, we’ll examine how two traditional tools of the trade perform when it comes to detecting synthetic identities.
Why it’s hard to detect synthetic identities with existing tools
Synthetic identities can be tough to detect, especially with traditional tools. Unlike identity theft in which a legitimate identity is used for fraudulent purposes, synthetic identities are purpose-built. The chances that a synthetic identity will fly through legacy identity-verification tools is high—we see it often. A criminal who has created an identity will know the name, date of birth, address, Social Security number, and other details a traditional fraud detection approach checks for. Unlike a stolen identity, there’s no need to modify an address or phone number; the one on record is the one a criminal is likely to use.
It’s also not effective to solely rely on a credit score to detect synthetic identity fraud. Traditional credit scores, particularly those based on trended or time-series data, are a great tool for assessing credit risk, but identity-related data assets are needed when it comes to first-party fraud detection. Synthetic fraudsters start with non-existing credit and systematically groom the credit profile up through the credit spectrum, undetectable with traditional credit risk scores. One could say that many of these applicants are people that would be declined because they don’t meet your credit criteria, even though the credit profile they’re using appears like they would. Without purpose-built tools focused on synthetic, finding them is like searching for a needle in a haystack of legitimate customers.
Importantly, there’s a population of applicants with good credit scores that has a high likelihood of charging off, and exhibits patterns consistent with a synthetic identity. Finding them is the challenge—this means identifying the applicants who look like good consumers from a credit-risk standpoint, but that show signs of synthetic fraud. It’s imperative to identify these at the time of booking, before the balances build and losses occur.
What we’re doing about it
The vast majority of your applicants are good, and the last thing you want to do is incorrectly flag a good applicant as a potential risk. TransUnion has validated our synthetic score offering with a number of customers with promising results. By taking a different approach to analyzing consumer data, enabled by unique TransUnion technology and our CreditVision suite, we can identify synthetic identities early in the credit building process, at the time credit is applied for and with extremely low false positives. Early analysis shows an ability to isolate less than half of a percent of a typical applicant pool that has a rate of charge-off well in excess of normal rates and showing signs of being a synthetic identity.
As synthetic fraud grows, we’re excited to be able to leverage our unique content and analytical power to deliver a simple score that effectively detects this threat.