Credit scores are created by for-profit institutions whose mission is to maximize profits. The data and the algorithm that go into the score are managed by opaque, unwieldy institutions. A 2004 study revealed that 79% of credit reports contained errors; 25% of these mistakes were serious enough to result in a credit denial. More than half of all credit reports contained outdated information or information belonging to someone else. You might think that because the rich use credit so much more, they’d be the ones mainly affected by these errors. In actuality, the poorer you are, the more likely your credit agency is to make a mistake that influences your rating. There is no accountability or transparency for the CRAs and no incentive for them to provide accurate information.

There are also problems with FICO, a for-profit institution, creating algorithms. Because an algorithm is just a mathematical formula, FICO has to use restrictive patents so it can make money from the algorithm. No one can know the formula, lest they replicate it, causing FICO to lose profits. Hence, it is in FICO’s best interest to keep it secret. When individuals are told “their scores,” these usually aren’t the FICO scores that are used in most lending decisions. It’s in the score-providers’s interest to obfuscate this fact so that consumers continue to buy their scores.

Besides the fact that the algorithm created for a profit, there are methodological issues. First, the model omits variables that are critical to repaying debts on time, such as income, assets, and credit terms. The model also assumes that borrowers are “statistically independent” of one another, meaning that one’s default depends only on one’s individual characteristics and is unaffected by the defaults of others. However, defaults are oftenstrongly correlated (e.g., a foreclosure depresses housing values across a neighborhood, making property owners de facto poorer, which increases the probability of further foreclosures). Finally, the model assumes that consumers’ riskiness or creditworthiness determines their credit scores. But in practice, low scores result in worse loan conditions that in turn increase their riskiness. Thus, the credit score functions more like a self-fulfilling prophecy than an “objective” prediction.

As with other components of the debt system, credit scoring negatively impacts people of color at a disproportionate rate. About 42% of Latino/as and nearly half of Black people in the United States have credit scores under 660, compared to just under 20% for Whites. While the median credit score for Whites rose from 727 to 738 during the 1990s, it decreased for Blacks from 693 to 676, and for Latino/as from 695 to 670. Just like the racial wealth gap (seeIntroduction), the racial gap in credit scores has worsened, not improved, over recent decades.

Finally, misuse of credit scores is now rampant. Using them to determine access to housing, employment, or health care is absolutely deplorable, even more misguided and harmful than using them to determine simple access to credit. These decisions have grave consequences for people’s lives; housing, health care, and the means of subsistence are fundamental human needs that should be available to all, not granted or withheld based on a person’s credit history.