Even when name matching is done by KYC solutions and software, factors like culture, and geography. And media reporting poses difficulties. This is the first article in our series on anthropolinguistic insights. We demonstrate how GRID data and Review screening function in relation to these larger tendencies.
Know your customer (KYC solutions) risk screening is based on a subliminal theme: names. The most distinctive identifier of an individual is their name. Which is frequently cited in negative linked media. Although Moody’s Analytics KYC solutions take into account a number of identifying factors. Such as a person’s location and date of birth, name matching is the key to producing precise warnings.
It may seem straightforward, but it’s not. Names can be very intricate. In a global dataset of names collected from sources such as government watchlists and negative media from more than 240 nations — in more than 70 languages and multiple scripts — name matching is particularly challenging. When risk-relevant individuals and organizations are named, some of whom might prefer to go undetected, the problems increase.
Moody’s Analytics KYC solutions have assembled a group of cross-functional screening experts to oversee this procedure. To solve name-matching hurdles and produce sophisticated results, this team blends qualitative ethnolinguistic skills with specifically designed quantitative data science methods.
Not just a game of numbers
People are not simple, however, the software is. It’s not easy to build software about people for people. Several factors need to be taken into account in order to accommodate the complexity of our existence, including our identities.
There are numerous naming systems used around the globe to arrange personal, kinship, and cultural names. Depending on the person’s gender, surnames or family names may appear at the start or end of a full name. For instance, in Russian-speaking societies, women’s surnames frequently end in “a,” although their brothers’ surnames do not.
Arabic naming customs pose a special problem for name matching since the inclusion or omission of honorifics or other name elements varies greatly and name order generally does not map exactly to that of other cultures. The diaspora of Arabic names and the irregularities in name recording are related.
In countries with comparable naming practices, name order might be crucial in differentiating names. For instance, Hispanic names typically have the paternal surname first, followed by the maternal surname, but the reverse is typical for Lusophone names.
Some societies pass down meaningful or symbolic names through the years, resulting in names that are widely used. Vietnamese middle names, for instance, frequently match with birth order, providing less identifiable information than a name component would otherwise. These practices make it difficult to determine the optimal way to use distinctive name components in software development without clogging screening results with false positives.
For identification of disambiguation among common names, Moody’s Analytics offers three major methods:
Particularly among premium content profiles enriched with entity-based research, GRID profiles uphold high standards for identifying information, including date of birth, residence, and aliases.
Alternate identities are analyzed and ruled out using filters that can be customized for alias, address, and date of birth.
To limit notifications to only high-confidence matches, a review does extra post-match analysis on GRID content profiles.
Names can become trendy due to well-known people, migration waves, or birth booms. Names can lose a lot of their appeal just as soon. The infamous “bad men” of GRID today might prevent their identities from becoming more well-known in the future. Some countries may restrict some names because it just takes one poor reputation to generate problems for their namesakes.
When transliterating a name, recording it from one media to another can easily produce problems. Who conducts an audit of an entity’s journey from A to B? However, depending on a signature to spell out a name can also be troublesome. audio-to-text transcription introduces several new issues.
Fortunately, we don’t frequently need to rely on handwriting to spell a name thanks to contemporary word processing. However, there are still disparities between and even within nations in the use of digital records, literacy rates, and computer access. Optical character recognition, which provides generally accurate but occasionally erroneous readings of printed documents and letters, may be used for name recording. Even in areas where digitalization is common, the software can cause mistakes, such as “correcting” “Kara” to “Lara” or “Mark” automatically.
Disambiguation is less straightforward than it seems when identifying a typo across languages and scripts. Such unpredictability is increased by the variety of keyboard templates. Like French and English, languages that share the same alphabet may have their own keyboard layouts, although logographic scripts may fall victim to the “Wubi Effect,” in which a group speaking the same language frequently uses multiple competing keyboard layouts.
Our Review program integrates targeted information on acceptable name variants to support transliterations, as well as frequency data on misspellings to evaluate potential typographical problems and prioritize pertinent notifications.
Effects of geography
Different regions have different languages, scripts, writing requirements, and cultural constraints, which all change through time. Unexpectedly, political power affects how names are spelled, documented, and transliterated. As a means of promoting social solidarity, the Russian language and Cyrillic alphabet were mandated in the Soviet countries.
In other instances, the issue is cultural resuscitation. After Ireland gained its independence in 1922, the nearly extinct Gaelic language was restored. A rule that requires the use of Gaelic place names on official maps and road signs in some sections of the country’s west only went into effect in 2005.
When two people have names that don’t match—and therefore shouldn’t match—name matching becomes considerably more challenging. Although some people would choose to remain anonymous, name changes can also have other benefits. In some places, legal name changes are more frequent than in others, and they are frequently governed by laws. In 2005, the supreme court of South Korea permitted name modifications under specific circumstances, such as when the name is similar to an objectionable phrase or is the name of a well-known criminal. A person who has a criminal record or a history of bankruptcy is prohibited from changing their name under this law.
A person may use a different name in another place because of cultural pressure, a desire to assimilate, or possibly to make it simpler for other people to pronounce them.
To enable the identification of risk entities, GRID profiles capture all possible aliases as noted in negative media coverage. Our high-risk premium content packages, such as Politically Exposed Persons (PEPs) + Family or Close Associates, Iran Connect, and Sanctions Connect, feature additional research. In order to find high-risk sanctioned entities, review screening also contains essential watchlist logic and optional OFAC-search type algorithms.
Overcoming the difficulties
Our KYC process and solutions assist you in making more accurate, better judgments faster and lower barriers to dealing with people; they are meant to support operational resilience for our clients.
By examining GRID data, our analysts can spot changes over time as well as reveal trends in name variants among areas. And the GRID software is developed using these findings.
To handle challenging, cross-cultural realities, GRID name-matching teams continuously develop new software and skills. For a flexible and consultative approach to risk-screening best practises, we depend on software intelligence and development, strong governance processes and support, and ethnolinguistic experience and insights.