A Brief History of Credit Scoring
* The article was initially published on our blog.
From paper “cuff” to face recognition the approaches to credit scoring changes drastically over time. Armada Labs recaps the milestones of credit scoring development, explaining what “ultimate” FICO means and how technology innovation impact the way modern companies underwrite a loan today.
For me, as a twenty-something-year-old, it is hard to believe that once it was impossible to buy things you like right away if you don’t have enough cash on you. That you were forced to go through cumbersome paperwork and then wait like a week or more until your loan application will finally be approved. That people with little to no credit history were left with no choice but to take rip-off “until-payday” from shady street lenders instead of submitting their applications to accredited financial institutions.
It all seems like a bad dream today when the financial industry is awash with aspiring entrepreneurial companies taking that blend of finance and technology, drawing on new types of digital data to extend credit box for those underserved and make real-time credit decisions based on enhanced analytics.
First Credit Bureaus
Well, as you may notice, the things were not so bright all the way. Back into the 20th century, the early American credit bureaus were local operations that scoured newspapers for information including notices of arrests, marriages, promotions and more. They collected all sorts of dubious stuff about people’s marital troubles, sex lives, and political activities and then sold it to merchants to refer to before granting credit. The problem was that nobody has been verifying the information or the sources it is coming from. How do you like this data-mining?
In 1956 engineer William Fair and mathematician Earl Isaac have sprung up a little of science into the process. They came up with an algorithm that parses the data looking for patterns to make a prediction about borrowers’ behavior. It turned out that people who are responsible in their home and the way they drive their car will probably maintain their credit. And those who drive fast on the highway probably drive their credit fast.
Thus, Fair, Isaac and Company became the first consultancy, whose product was a two-page scorecard, made of cardboard, provided to banking and retail customers with all the information they needed to make a credit decision within 48 hours.
Regulatory Oversight
Along with the rapid growth of Fair, Isaac and Company, another three major consumer credit bureaus in the U.S. were gaining traction. Their names are widely known until today and they are Retail Credit Company (now Equifax), Experian and Trans Union. Together these companies held records on millions of Americans and cheerfully shared them with any buyer. Their power and spread very soon attracted regulators’ attention and amid a debate that presaged today’s fights over data privacy, Congress held hearings into the matter.
In 1970 the Fair Credit Reporting Act (FCRA) was enacted to promote accuracy, fairness, and the privacy of personal information assembled by credit reporting agencies. By this passage, consumers received the rights to see and make amendments to their personal files while credit bureaus were obliged to report information only to those with a legitimate purpose.
It was followed by the Equal Credit Opportunity Act (ECOA) of 1974 that made it illegal to discriminate applicants on the basis of sex or marital status. In 1976 the list was broadened with the consideration of race, religion and several other characteristics.
Although regulation restricted the activities of credit bureaus, it also galvanized the scientific approach towards credit scoring making it not about individuals but about algorithms.
Credit Scoring As We Know It Today
The biggest revolution in the credit-scoring that made it look more of what we expect of it today came in the 1990s. Working closely with Equifax, Experian and TransUnion, in 1989 Fair Isaac and Company unveiled the first consumer-credit score — a number between 300 and 850, where higher scores indicate a better credit rating. Known as the FICO (for Fair Isaac Corporation) score, it rapidly became the standard for American lenders.
Fair Isaac and Company was no longer developing custom algorithms for lenders by mapping heir past customers’ attributes onto future ones. Its new FICO model assigned a three-digit number to every individual in the system based on the swaths of data from the major credit agencies. Although the exact formulas for calculating credit scores are secret, it is known that a FICO score results by applying a different weight to five core bits of financial information including payment history, the total already owed, length of credit history and two scores for the mix of credit: cards, shop accounts and mortgages, and applications for new credit.
Today lenders, such as banks and credit card companies, use FICO credit scores to evaluate the potential risk posed by lending money to consumers. Thanks to FICO, which now underlays 90% of credit decisions in the U.S., credit became widely available and less expensive for many consumers.
However, credit scoring techniques never stop evolving. With the emergence of alternative data sources and self-learning algorithms, we face an even greater number of ways to underwrite a loan that varies from country to country. For example, in China, where digitalization of society has reached the highest heights, the new system of “social scoring” was developed. The benefits of a good “social score” go far beyond borrowing, they include easier visa applications, lower rental deposits and even better placement on dating sites.
But in smaller, poorer countries with little financial infrastructure, credit-scorers have limited data to work with as people usually do not have a bank account. They have a mobile phone which could help them access financial services, though. Many financial services companies operating in these markets do not make underwriting reliant on traditional credit scores but seek to increase consumers’ willingness to pay by taking advantage of their desire to access the phone.
PayJoy, a San Francisco-based startup that operates in Mexico, India, Indonesia, Nigeria, Kenya, and Guatemala enables consumers with no bank accounts or formal credit history to purchase smartphones on installment payments and get cash loans. It does so by turning the smartphone into collateral through software that locks the phone when payments have not been made.
Conclusion
The group of startups like PayJoy has spied an opportunity to grow and expand fair credit to those underserved. Though their target audience usually lays outside America, they stay free from race, religion or gender discrimination. Still, they take into account some attributes that would set a Western data-protection regulator’s blood pressure soaring. From their point of view, good borrowers include those who put both first and last names in their contacts, those whose travel and location follow predictable patterns, and those who communicate regularly with a few contacts. Does not it resemble something we discuss at the beginning of the article?
Yet the difference between robust credit-scoring in the rich world and the novel methods now used in emerging markets is likely to be temporary. Of course, these exotic correlations have no power in comparison to the predictiveness of actual financial data, but it is okay for now. By the time people out there accumulate records and gain a certain understanding of financial services their applications will be considered in the developed countries.
Thanks for reading! Please give a clap to the story if you enjoyed and check out what Armada Labs does 😊