Knowledge Center: Barry Graubart on Complexity of Identity Resolution and Anti-Money Laundering
By Barry Graubart, VP of product management, Alacra
In the retail world, identity resolution and data cleaning applications (such as fraud detection) have the luxury of fairly comprehensive data. Right there in your customer and employee databases you have name, address, phone, SSN, date of
birth and many other values to match against.
However, when comparing customer records to Anti-Money Laundering (AML) or Terrorism watch lists, banks and other financial institutions frequently have very limited access to information (on the watch list side). Terrorists, money launderers and drug traffickers don’t often provide a SSN or a street address. Instead, AML pros are limited to matching a name and, at best, city or country. Further complicating matters is the fact that to Westerners the cultural variations in name spelling make matching difficult, particularly with Asian and Middle Eastern names.
As an example, the Libyan head of state’s name can be spelled the following
ways:
Muammar al-Qaddafi
Moamar El Kadhaafi
Mu’ammar Al Qathafi
Mu’ammar Al Qathafi
Mo’ammar Gadhafi
Moammar Khadafy
Resolving identities, when you factor in that each of these spellings is correct, grows exponentially more complicated. Toss in accidental error and intentional attempts to deceive and you begin to see how tricky AML compliance can be and why costs to North American institutions have risen 70% in the last three years.
Due to cultural differences, data entry errors or deliberate data manipulations by money launderers, finding matches within your data is often problematic or impossible with common search techniques. Many methodologies used to identify individuals depend on finding matches within databases. This works fine in theory, but in reality, unless the data matches perfectly, these methodologies are ineffective and increase the rate of false negatives or incorrectly discarded records.
To catch a sophisticated criminal, you need to implement sophisticated similarity search techniques to resolve multiple identities into one unified view.
The importance of having an effective identity resolution solution in place that can resolve all these complexities is vitally important because the consequences of failing to detect a money launderer are much, much higher than when a shoplifter lifts a pair of pants. In More Data is Better, Proceed With Caution, Jeff Jonas eloquently writes:
“The opportunity cost or consequence related to situational awareness is proportional to the size of the organization – bad decisions in the case of the manager of the barber shop produces one worst case outcome, bad decisions at Enron another, a country another, and mankind yet another, each with increasing consequences.”
Subscribe to this blog