New Credit Data Deserves Both Hope and Skepticism

In recent years, companies have been using new technologies and data to create alternative credit scoring tools. These are developments that those concerned with poverty and economic opportunity should be watching closely.

New credit scores might eventually help expand access to credit, but it۪s important to be wary of the hype. Although there are some sensible innovations on the horizon, the most ambitious new scoring methods are not yet well-studied, and they might even open the door for new forms of discrimination and abuse.

Across the financial world, companies are eying new data. Fair Isaac (FICO), purveyor of the industry-standard FICO scores, announced that it is beginning to experiment with alternative data sources, such as cable bills and utility bills, to help identify new creditworthy individuals. More ambitious still, a new breed of startup companies claims to analyze social networks, LinkedIn data, online quizzes, and other Internet data sources to evaluate new borrowers. “All data is credit data,” declares ZestFinance, exemplifying the spirit of these startups.

These announcements focus on the tens of millions of people who do not have enough financial history to generate a traditional credit score. Today, most credit scores rely on a narrow category of historical credit data such as past repayment behavior held by credit bureaus. According to the National Credit Reporting Association, as many as 70 million Americans do not have a credit score, or have a lower score than their full financial history would warrant.

This is a problem, because credit scores are important. “A good credit score can mean access to a wide range of credit products at the better rates available in the market, while a bad credit score can lead to greatly reduced access to credit and much higher borrowing costs,” explains the Consumer Financial Protection Bureau.

So, is new data actually poised to help score more people in a fair and accurate way? It depends.

There is some early evidence that bill repayment data such as cable bills, utility bills, and rental payments might help fairly and accurately score more people. This makes sense, because these records resemble the ingredients of today۪s credit scores. FICO claims that, by considering such data, where available, it can “reliably [score] 15 million consumers who do not have enough credit data to generate FICO scores.” More public study is necessary to ensure that this data yields fair results, and advocates must help ensure that consideration of utility data does not interfere with state laws designed to shield vulnerable populations. But there is reason for optimism.

On the other hand, many of the newest companies which often rely on social media, shopping, and other online data sources to evaluate creditworthiness deserve far more skepticism.

ZestFinance, one such business, uses thousands of points of data to try and improve or supplant normal credit scores, which it says will allow subprime lenders to offer more loans at lower costs.

However, there is no public evidence that the types of data used by ZestFinance and similar companies provide the predictive punch of traditional credit data. Traditional credit scoring models, though not free of problems, are well-studied and reasonably well-understood. The newest models are not.

Moreover, many new companies operate at the frontier of existing regulatory policy. It is not clear how fair lending laws, which prohibit discrimination against a loan applicant on the basis of race, sex, and other factors, will apply. (When a computer is set loose to learn from a large, diverse pile of data, it may discriminate against a protected class, even if its designers didn۪t mean for it to do so.) New protections for consumers may be needed to ensure that alternative data models do not discriminate against vulnerable communities.

In short, credit scores will be important financial credentials for years to come. New approaches to credit scoring could create opportunity for millions of people who would be better able to start a business, buy a home, or borrow money in emergencies. Advocates should be on the lookout for new uses of data that are both predictive and fair, but they should avoid getting caught up in the hype.

To print a PDF version of this document, click here.

Aaron Rieke is director of Tech Policy Projects at Upturn. He was co-author of the recent report Knowing the Score: New Data, Underwriting, and Marketing in the Consumer Credit Marketplace.

The views expressed in this commentary are those of the author or authors alone, and not those of Spotlight. Spotlight is a non-partisan initiative, and Spotlight’s commentary section includes diverse perspectives on poverty. If you have a question about a commentary, please don’t hesitate to contact us at commentary@spotlightonpoverty.org. 



If you wish to submit for consideration a commentary to Spotlight, please visit our commentary guidelines and submission page.