Skip to main content
Solving the “Black Hole Problem” with First-Party Fraud

At FICO, we have the privilege of working with the smartest, most innovative folks in the financial industry on the thorniest, most complex problems facing the industry. This work is challenging, rewarding, and frequently filled with fascinating conversations.

For your benefit, we will periodically reproduce these conversations in FICO blog posts. Today’s post is about solving the elusive problem of first-party fraud and features commentary from Julie Conroy (Research Director at Aite Group) and Jack Alton (CEO at Neuro-ID).

Alex: So here’s something, recently overheard at a banking conference, that I found interesting. An executive who owns P&L for a consumer lending product line at a large bank was asked about first-party fraud and he compared the problem to a black hole. What he meant was that even though first-party fraud has a very real impact on his balance sheet—everything from spending money to underwrite applications to assigning precious collections resources to treat delinquent accounts—it is nearly impossible to observe or measure because it looks virtually identical to the legitimate credit losses on his books.

Much in the same way that black holes are real natural phenomena, but are nearly impossible to observe because they don’t reflect light or give off any meaningful amount of heat.

I found this description compelling because the idea that something can be both massively influential and practically impossible to observe seems to boggle the mind, to such a degree, that even the smartest and most capable people seem to prefer to just ignore it altogether. Related—Albert Einstein, whose equations predicted the existence of black holes, found the concept so disquieting that he eventually wrote that the idea was “not convincing” and the phenomena did not exist “in the real world.”

In my experience, bank executives tend to treat first-party fraud the same way. Logically, they know it’s real and has a big impact on their business, but most actively choose to ignore the implications because of how hard it is to define and identify.  

Julie—I know you spend a lot of time researching this subject. Why is this such a hard problem for bank executives to wrap their heads around?

Julie: There are three primary reasons that synthetic identity fraud has been an elusive problem for quite some time:

  1. As you pointed out earlier, a lot of synthetic fraud gets written off as a credit loss—in the U.S., this is actually by regulatory mandate, which stipulates that never-pay-defaults are categorized as credit losses, even though many of them often turn out to have a synthetic identity component. This makes it a credit risk problem, but the challenge is that the fraud group is the team best equipped to detect synthetics through a variety of techniques. Thus, the age-old challenge of who is going to fund the business case has been one obstacle to progress.

  1. Synthetics are often really hard to detect—the organized crime rings behind much of this activity build the synthetic identities patiently over time, emulating the behavior of a genuine consumer that is new to credit or new to the country.

  1. Across the industry, we don’t have a consistent definition for synthetic identity fraud. When I did the research for the report I published last year on this topic, all of the issuers I interviewed had slightly different definitions for what constituted a synthetic in their FI (and a couple said that the definition varied within their FI as well, across different lines of businesses). Without common definitions and benchmarking, it’s hard to coalesce around common solutions. 

However, I do see a much greater willingness to tackle the problem now than I did four or five years ago. I think this is due in large part to the fact that the incidence of synthetics has substantially increased in the U.S. over the course of the past decade. This is due to the convergence of a perfect storm of contributing factors: loosening credit standards during the economic recovery, which made it easier for the newly minted identity to gain an initial foothold in the bureaus; the post-EMV fraud migration, as organized rings sought ways to backfill their revenues as counterfeit fraud was worked out of the system; scores of data breaches that provide those crime rings with an ample supply of data to use to fabricate the identities; and finally the 2011 randomization of Social Security numbers, which served to make it much easier to mint a new identity from scratch.

A key indicator that the problem is substantial and growing has manifested for a couple of the FIs I interviewed for last year’s report. Both were seeing substantial shifts in their delinquency curves that were only explainable by a big uptick in synthetic identity fraud. Skip-tracing during collections often leads back to a real person whose identity was used as part of the identity-compilation process, so the account is removed from the collections queue. Enough of these cases shift the FI’s delinquency curve. Another factor bringing the credit risk function to the table is that some FIs’ credit losses are worsening even though the credit models themselves aren’t degrading. This again points to synthetic identities as a culprit. So the good news in all of this, is we’re finally seeing recognition of the issue, and financial services firms investing time and resources in addressing it.

Alex: Returning to the black hole analogy, it sounds like even though we still can’t see it (or even agree on a precise definition of what it is) banks are feeling the impact of first-party/synthetic identity fraud in a more significant way and are taking steps to start detecting it.

Jack—I know Neuro-ID has been doing some really interesting work in this area. How are banks attempting to solve this problem?    

Jack: As Julie said, much of what we are discussing hinges on a better understanding of the problem, what it affects and the best techniques to detect it. 

At Neuro-ID, we have run into many of the same obstacles that Julie referenced. We also believe that there is hesitancy to address the problem due to how difficult it is to detect synthetic identities, as well as first-party fraud, both of which end up getting rolled into credit loss. The current blind spot at financial institutions (FIs) really comes down to understanding and acting upon the intent of their applicants.

The good news is that new advancements in real-time behavioral analytics allow FIs to measure friction in their customer experience, as well as identify behavioral markers of fraud and mal-intent within the new account opening journey. This new visibility enables FIs to both identify and treat synthetic identity fraud independently from credit loss.

Remember, black holes were a purely theoretical concept until earlier this year, when technology was finally able to prove what theoretical physicists have known to be true for decades.

At Neuro-ID, our goal is to use technology to better understand ‘digital body language’ and the intent of the applicant. We can prove, through the science and data, why customers are reacting the way they do and the intention of applicants as they interact online. In other words, behavioral intelligence is the new lens that makes it possible to see elusive synthetic identity fraud. And now that it can be clearly identified and isolated, businesses can better understand the problem, its impact on their P&L, and how to best treat the cause.   

related posts