FICO scores are objective, relevant and reliable

To build on two previous articles unpacking the current Mercator Advisory Group white paper Credit Scoring, Fintech, and Consumer Credit: Why AI Scoring Models Are Not Replacing the FICO Score, payment journal sat with Brian RileyDirector of the Credit Advisory Services Practice at Mercator Advisory Group to learn more about why industry-leading FICO credit scores are the most reliable measure of creditworthiness.

Fairness and objectivity in credit assessments

Financial institutions need accurate metrics to make decisions, control risk and assess credit quality. Since 1989, the FICO Score has used factual data to rank risk and relies on information provided by creditors. The underlying information comes from five data points: loan repayment history, amount owed, length of credit history, recency of new loan applications, and type of credit history. The FICO Score uses the accurate sources of information to provide an accurate, consistent, and fair measure that encompasses all facets of secured and unsecured consumer credit.

“The FICO score sticks to the facts that regulators rule. It doesn’t try to introduce casual or social elements. The score creates a relative ranking based on the risk of the account,” Riley said. “Regardless of the client’s background, a 660 means the same thing to any borrower anywhere in the United States. So a 520 FICO score and an 800 FICO score.”

FICO’s approach has two key benefits. First, the data used to calculate scores is straightforward and regulated to ensure that it is inherently unbiased towards individuals or groups. Second, the calculation of FICO scores has been tried and tested for decades and is transparent. FICO’s transparency contrasts with newcomers to the credit scoring industry like UpStart, which uses AI-powered systems that are effectively black boxes when calculating credit scores. Such scores can arouse suspicion because of their somber origins.

Machine learning shows promise in consumer credit, and there is evidence that artificial intelligence is evolving into space. While there may be substance, the models rely on hype or unregulated data that could be misleading or unfair. Other models consider data used in calculating FICO scores, but attempt to go beyond traditional boundaries with data elements such as college education, social media presence, and past purchases. These models aim to open the underwriting gate and bring in the credit-unseen, the underbanked, or the credit-damaged. However, these plans risk introducing bias and creating a credit rating system that is impossible for people to understand and even harder to justify.

A transparent credit check system is essential. If a loan application is rejected, the applicant requires an explanation. Not only is this a good deal, but it is also required by various regulations such as Fair Lending and Fair Credit Reporting. Transparency is a fundamental part of the FICO score, but many alternative models fall short.

Credit Scoring Bias

In recent months, the use of certain alternative data in credit assessments has drawn resistance from policymakers. These events triggered the introduction of a new one invoice in the House of Representatives asking the Consumer Financial Protection Bureau to assess the use of educational data by consumer lenders in their underwriting processes, to publish that assessment, and to present its findings and recommendations to Congress on how to address potential inequalities.

Unlike some fintech AI models, the FICO score has met fair lending requirements for decades. Fair credit regulators have determined that the FICO score has no prediction bias versus protected classes. When comparing people with the same probability of repayment or default, the model ranked people in these protected groups no worse than people in the general population. In an environment where racial justice is paramount, credit ratings that prove fair over decades should be the gold standard.

A lack of transparency when calculating creditworthiness is a problem

As mentioned, companies like Upstart use machine learning algorithms that are difficult for mere mortals to understand. Highly flexible machine learning algorithms often have limited visibility. It is often extremely difficult to understand a variable’s contribution to a prediction, how the variables interact with each other, and why the algorithm may have considered the variable important. When these algorithms are particularly complex, the term “black box” indicates that the algorithm lacks clarity and the predictions are unreasonable or inexplicable.

Given that fair credit laws and federal regulations require a lender to clearly explain loan denials, companies that use machine learning algorithms to generate credit scores can find themselves in a precarious legal position. The inherent weakness, lack of transparency, and legal ramifications could be why the stock prices of companies like Upstart have plummeted recently. This indicates a lack of market confidence in the underlying business models.

Credit Scoring and the Inevitable Recession

With the recession approaching, businesses need to rely on reliable and innovative credit scoring. FICO has been in business for decades and has established an enduring, pervasive risk assessment metric. Emerging companies don’t yet have data on how their model performs in a recession, so they’re virtually untested in such environments. Now is not the time for a bank to base its credit risk assessment on emerging, untested models.

Additionally, FICO is an industry leader, being the first to bring to market tools that subtly incorporate additional data into their models. To prevent lenders and consumers from taking on more risk than they can handle, the FICO score is slowly expanding to allow relevant data points to supplement the data provided for the three major credit bureaus (Experian, Equifax, and TransUnion).

“There’s going to be a horizon where the change happens and don’t expect it to be rapid, but expect it to be very thoughtful,” Riley said.

A recent example of the volatility of alternative scoring can be seen in the recent Securities and Exchange Commission (SEC) filings of Oportun, a fintech lender that uses a proprietary rating to target the unrated population. in one current investor report, the company notes that it has helped create credit histories for 1 million people through its artificial intelligence scoring model. While that’s an exciting claim, it’s interesting to note that the APR for lending products is at the higher end of the spectrum, with personal loans having an APR of 32.3%, followed by secured personal loans at 29.1% and credit cards at 29.8%. These high interest rates are important facets of their credit acceptance model to account for non-scoring and indicate the risk associated with AI scoring. Unlike the credit card APR at Oportun, the Federal Reserve reports that the average APR for interest-bearing accounts as of May 2022 was 15.13%, nearly half the rate charged by Oportun.

High interest rates are required when looking at loan defaults. in Oportun, Annualized net charge-off rates for the six months ended June 30, from 7.5% in 2021 to 8.8% in 2022, and now that the United States faces the threat of prolonged inflation, loan losses are trending towards the company’s highs which reached 9.8% in 2020.

Riley provided the example of rent and mortgage payments in different parts of the country to illustrate the FICO score’s inclusion of relevant data. A tenant in Chicago and a homeowner in Sioux Falls may receive different credit ratings, but both can demonstrate responsible, timely payments related to their home. These and other similar factors appear in different versions of the FICO score:

  • FICO 8: The most widely used version of the standard credit score model, using the five primary metrics as the core rubric for credit scores from 300 to 850.
  • FICO 9: This release includes adjustments to the handling of medical collections accounts, rental history, and third-party collections.
  • FICO 10/10T: A more predictive version of the FICO score that uses credit activity over the past two years to account for potential future risks.
    • FICO 10T: The T stands for “trend data‘ included in this score version.
  • UltraFICO: An opt-in lending model that helps individuals improve their FICO score by evaluating personal banking information and behavior, including cash on hand, frequency of transactions, and positive balance history.
  • FICO XD: An alternative credit score created in partnership with LexisNexis and Equifax that can rate borrowers with little to no credit history based on utility, cable, and phone bills.

By progressively and strategically applying a mix of data to strengthen risk profiles, different versions of the FICO Score can bring new consumers on board without sacrificing the regulatory oversight that makes it such a solid scoring standard. Rest assured that the FICO Score will evolve safely and responsibly, maintaining its essential integrity while reflecting the realities of the modern world.

Comments are closed.