Psychometric credit scoring uses personality tests to gauge creditworthiness. Could it come to the U.S.?
Do you make your bed every morning? If so, you’re more likely to be approved for credit.
Credit card issuers and lenders in the U.S. typically don’t ask prospective borrowers these kinds of questions when determining creditworthiness. But a select few of the 3 billion or so people outside of America who have no bank accounts or credit files are beginning to see such queries on loan applications. It’s not just oddly placed small talk – the questions are meant to find out what kind of person the borrower is and use that information to predict his or her ability to repay.
Welcome to the world of psychometric credit scoring, a credit risk assessment model that evaluates personality traits such as confidence, conscientiousness, community perception and, yes, bed-making acumen. Psychometric scoring models also can mirror more traditional personality tests that assess traits such as openness, extroversion, agreeableness and neuroticism.
Psychometric credit scoring is a relatively new concept, but has proven effective in expanding credit markets around the globe without increasing lending risk. Though it has yet to affect the financial lives of Americans, U.S. credit scoring heavyweights FICO and Equifax have found success with psychometric testing abroad.
Do you save enough? And are your neighbors lazy?
Psychometric credit scoring was pioneered by Entrepreneurial Finance Lab (EFL), a company that began over a decade ago as a Harvard research project on entrepreneurial finance in emerging markets. Over the years EFL has expanded its reach to about 1 million loan and credit card applications worldwide and secured partnerships with Mastercard and major banks and retailers in Latin America, Africa, Asia and Indonesia.
EFL’s questionnaire – which can be delivered via desktop, mobile apps and even text message – varies wildly in terms of its apparent relevance to personal finance. One question asks how long you could get by if you lost your main source of income, while another gives you a choice between 50 gold coins today or 100 coins one year from now. Another item asks you to rate the work ethic of other people in your community by moving a slider between two different images – one of a cartoon man hammering a nail and another of the same man taking a nap.
Still another asks where you fall along the continuum of leaving your bedroom in disarray in the morning or making hospital corners, as seen in the example below.
What does any of this have to do with your ability to repay a loan or a credit card balance? EFL Product Marketing Manager Amie Vaccaro said the firm uses more than a decade’s worth of loan performance data to measure how certain personality traits correlate with credit risk.
“We don’t go in saying that we think people with high levels of confidence are going to have lower credit risk,” Vaccaro said. “We look to see what’s proven out in the data.”
Vaccaro used “happiness” as an example of a trait that could provide insight into a person’s creditworthiness. She said people with moderately high levels of happiness are found to be low-risk, while others who are extremely happy tend to be riskier. Similarly, too much ambition – a quality many entrepreneurs would be proud to flaunt – could be indicative of high credit risk.
The EFL model also takes into account the rate at which you answer the questions. You don’t want to be too fast or too slow – the former indicates you’re just whizzing through without thinking about the questions, while the latter suggests you might be cheating.
“We look at a lot of different things to come up with an assessment of these traits,” Vaccaro said. “It might be more how you answer a question, as opposed to what your answer is.”
FICO, Equifax embrace psychometrics
If assessing a consumer’s creditworthiness based on happiness or ambition seems far-fetched, consider that none other than credit-scoring giant FICO is now marketing EFL’s model to major banks around the world. The two firms in April announced Russia’s Sovcombank was using psychometric scoring to extend credit to young consumers.
David Shellenberger, senior director of scoring and predictive analytics at FICO, said psychometric testing is valuable in areas where consumer data is sparse or credit bureaus don’t exist.
“What we like about psychometric scoring is that we don’t have to rely on any other type of data asset,” Shellenberger said. “It’s all consumer contributed. In certain circumstances, we’ll augment that data with other types of data sources to create a safe and reliable score.”
Shellenberger acknowledged that traditional and alternative data are stronger than psychometrics in terms of credit risk assessment. However, psychometric testing holds an accessibility advantage over the other methods.
“We can get psychometric information from anyone willing to take the survey,” he said.
Credit bureau Equifax has also tested how EFL’s model could help Latin American banks extend credit to small businesses. A pilot study showed blending psychometrics with traditional data from Equifax could increase lending by 140 percent while maintaining default rates, or reduce default rates by 50 percent while maintaining acceptance rates.
Will psychometric testing ever come to America?
At present, EFL has no plans to market psychometric credit scoring in the U.S. FICO estimates there are at least 50 million American “credit invisibles” – consumers who can’t get a credit score under the firm’s traditional model.
“That’s just so small compared to the 3 billion people globally who are unbanked or underbanked we’re looking to serve,” EFL’s Vaccaro said.
And EFL wouldn’t be the only game in town with regard to scoring the unscorable in the U.S. Shellenberger said his company’s FICO XD score, which incorporates alternative data, can help extend credit to about half of the unscorables in the U.S.
EFL also has company on the psychometric testing front. U.K.-based Coremetrix gleans psychological insights from an image-based test that looks similar to less-sophisticated pop culture/personality quizzes seen on social media sites.
Coremetrix’s test won’t tell you what “Full House” cast member you are, but it has been proven to reduce default rates by 23 percent, according to a 2014 report by Mastercard.
The company has operations in several countries across Europe, Asia and Latin America, but it doesn’t have a U.S. presence. But Clare McCaffery, managing director at Coremetrix, said the company has plans to take its model to the U.S., albeit not within the next two years.
“Our current focus is on territories where there is limited bureau information, such as India,” McCaffery said. “There are also fewer barriers to entry in these markets, so it makes sense right now. Our mission is to facilitate banking for all and this includes consumers in the U.S.”
Of course, any nontraditional credit-scoring system will have to pass muster with the Federal Trade Commission. There is concern that psychometric testing can’t generate legitimate reasons for denying credit to a consumer – a requirement under the Fair Credit Reporting Act. EFL’s Vaccaro said it doesn’t make sense for her firm to “jump through the hoops” of working in the U.S.
“We find that regulators internationally tend to look favorably upon what we’re doing, so we’re grateful for that,” she said.
If psychometric testing ever finds favor with U.S. regulators – and if the companies who use it see value in the American market – “credit invisibles” could become an outdated term. And for many people, the credit application process would become as simple as a taking a BuzzFeed quiz.