When it comes to school, and especially higher education, everyone wants extra credit — especially when it comes to paying for everything from books to room and board. The costs of going to school, especially for grad school, are going nowhere but up.
To that end, SelfScore, which bills itself as an analytics-based lender focused on financial services provided to international students, last week debuted its Achieve Card in partnership with Mastercard. The firm said that this is the second card in what is designed to be a series of cards that are slated to fill a hole that is left by the “traditional” credit scoring model based on FICO.
In a release detailing the launch, the firms noted that the 1 million international students who are currently enrolled in universities based in the United States contribute as much as $32 billion annually to the domestic economy through tuition and by paying everyday living expenses (of that tally, $9 billion is paid out for tuition alone).
In an interview with Karen Webster, SelfScore CEO Kalpesh Kapadia said that the new card product is one that helps give credit to students who do not have a credit history and where there is in fact no way to validate that they have a credit history.
As for the features of the card itself, SelfScore has said that there is a 0 percent intro APR on all purchases for the first six months, along with 1 percent cash back on all purchases. Initial limits of $1,500 can be increased to $5,000. And for international students, the added attraction and benefit comes with the fact that there are no foreign transaction fees, no security deposit required and no Social Security number required.
Why delve into this untapped pool? Kapadia stated that “before the advent of the internet and the computer, and platforms and social media,” certain technological and stringent criteria for lending had been “entrenched in the system.” Banks, he said, “are reluctant to use anything new .… It’s the old adage from the tech industry that you don’t get fired for buying from IBM or Cisco.”
The same holds true for banks, where in the risk department “you don’t get fired for using FICO.”
In addition, he noted, assets as traditionally scored through certain channels such as FICO and Experian are done so via tight bands, which can make credit even harder to come by. He said that SelfScore “talked to major banks [and other firms] such as Bank of America, American Express, Citi and Capital One, and we realized that the best way to approach this is to productize around this.”
For one thing, said Kapadia, after the terrorist attacks of September 11, the United States government started viewing foreign students with greater scrutiny and issuing Social Security numbers. The trend has been for “being vetted thoroughly,” he said, by the university, financial institutions and embassies, which will want to see if an individual has any criminal background, along with a wealth of data and information that can be gleaned from social media. This sets the stage for machine learning to be employed by SelfScore in helping creditworthiness through various conduits.
As for the parameters governing international students applying for credit lines in the U.S., Kapadia noted there is an age threshold, where a student must be at least 18 years of age, and they must be of the class (and enrolled) that is being indicated in regards to a school, which school must be accredited.
Beyond those initial hurdles, said Kapadia, the question is “do you have an identity that can be verified?” One way to verify that identity is through a Social Security number. In case the applicant does not have a Social Security number, he continued, there is also the ability to use a passport or a driver’s license — and, he noted, the passport is relatively standardized across nations, while the passport numbers themselves can be ascertained through databases.
The typical application has seven or eight data points, including names and date of birth, tied to the most basic building blocks of information that can be checked via APIs to see what has been reported and self-reported.
Kapadia noted that there is also tracking tied to the stability of identity. A key question to answer is whether someone can be reached, through physical mail, phone or email address. These data points are also checked through APIs (checks and funds will not be disbursed, for example, to a post office box address.)
Finally, the executive said, machine learning helps ascertain the applicant’s ability to pay. For example, Kapadia said, “are you able to make minimum monthly payments” on the credit that is being extended?
This includes having additional financial resources from which to draw if needed, say, with parental help. The firm also deploys methods to find out employability and earnings potential upon graduation.
“You are betting on the potential rather than the past” in the end, he noted.
These above data points and information are fed to an algorithm, and the applicant receives a score, said Kapadia, with the same scale as would be seen with a FICO score.
With the establishment of a credit line and then subsequent payments monthly showing the construction of a credit history, across, for example, Credit Karma, Kapadia said that there is a 92 percent correlation between the score that SelfScore gives out and the FICO score that is received. As for paydown behavior, Kapadia said that “it is 70/30, as 70 percent [pay down], 30 percent revolve.”
As for the vagaries of taking on debt, Kapadia stated that the interest rates are competitive even for prime customers — much less no-file customers. When queried about default rates, Kapadia said that rate as measured in units is slightly less than 1 percent and in dollars is slightly over 1 percent.