Access to affordable credit is essential to economic security and prosperity. Affordable credit provides a safety net when unexpected expenses arise and enables generational wealth-building through homeownership. In 2021, the Federal Reserve Bank of Minneapolis created an interactive data tool to better illustrate credit conditions in the seven-county Twin Cities area of Minnesota.1 The Twin Cities Community Credit Profile tracks key indicators of credit access over time and by ZIP Code, making it easier to measure residents’ participation in the credit economy. In an extension of our work with the tool’s dataset, we analyzed credit access at the census tract level using two measures: the share of adults who are in the credit economy and the median credit scores of those individuals. We find that neighborhoods with higher shares of residents of color tend to have less access to credit. Even after adjusting for household income, neighborhoods where the largest racial group is White residents tend to have more credit access and higher credit scores than neighborhoods where the largest racial or ethnic group is Asian, Black, or Latino/a.2 In fact, neighborhoods of color tend to have median credit scores that are lower than what their median income alone predicts.
Examining race/ethnicity and participation in the credit economy
One way of understanding the availability of credit at the neighborhood level is to look at what we call the credit-inclusion rate, or the share of adults with a credit file and a valid credit score. Overall, 94 percent of adults in the seven-county Twin Cities area have a credit file and a valid credit score. However, at the neighborhood level, which we define as the census tract level, there is wide variation.
Panel A of Figure 1 shows credit-inclusion rates in the Twin Cities area divided into quintiles, where the first quintile suggests the lowest credit inclusion and the fifth quintile the highest, and Panel B shows the largest racial or ethnic group at the census tract level.3 Minneapolis appears at the maps’ centers, with St. Paul slightly east of the centers. Panel A shows a generally higher level of credit inclusion in tracts outside of the two central cities. In Panel B, we see that tracts where the largest racial or ethnic group is a population of color are in parts of northern and southern Minneapolis, the eastern two-thirds of St. Paul, and the cities of Brooklyn Center and Brooklyn Park, which are located just northwest of Minneapolis. There is an overlap between neighborhoods where people of color live and neighborhoods with lower participation in the credit economy as defined by the credit-inclusion rate.
Participation in the credit economy is lower in neighborhoods with more residents of color
Loading map 1a...
Loading map 1b...
Next, we directly examine the relationship between credit inclusion and where people of color live. See Figure 2. Census tracts where Asian, Black, or Latino/a is the largest racial or ethnic group are more concentrated in the first quintile of credit inclusion. In fact, all three census tracts where Latino/a is the largest racial or ethnic group are in the first quintile. In other words, they tend to have lower credit inclusion than majority-White tracts.
Loading chart 2...
Differences in neighborhoods’ median income levels are one important potential explanation for the patterns in figures 1 and 2. There are many ways an individual could establish a credit history, one of which is opening a credit card for the first time. While there is no income requirement for opening a credit card, credit card issuers may use income as a proxy for the ability to pay the debt. We find that while there is a lot of overlap in the income distributions between census tracts with above-median and at-or-below-median credit inclusion, above-median-credit-inclusion tracts tend to have a higher median household income. In fact, without adjusting for income, a 1 percentage point increase in share of White residents in a given tract is associated with a 1 percentage point increase in the likelihood of that tract having an above-median credit-inclusion rate. Taking income into account, the increase in likelihood falls to 0.4 percentage points. As a result, while income can explain a large amount of the variation in credit inclusion by race or ethnicity, a small disparity remains even after holding income constant.
Credit scores are lower in neighborhoods with higher shares of residents of color
Credit scores are a primary tool that lenders use to determine who has access to credit and on what terms. Because they reflect underlying economic disparities, credit scores may be imperfect indicators of borrowers’ likelihood of repaying their debts. However, since the credit market relies on these scores, understanding how they are distributed across populations can inform efforts to build a more inclusive economy. Because of perceived risk, people with lower credit scores are more likely to receive higher-cost loans, if they can even access credit.
The credit rating agencies generally consider credit scores of 800 and up “excellent,” 740 to 799 “very good,” 670 to 739 “good,” and 580 to 699 “fair.”4 Overall, 13.8 percent of census tracts in the Twin Cities area have a median credit score of excellent and 65.1 percent of census tracts have a median credit score of very good. Only 2.0 percent of census tracts in the Twin Cities area have a median credit score of fair. Overall, the median credit score in the seven-county Twin Cities area is 772.
To better understand whether lenders will perceive residents within different neighborhoods as high or low credit risks, we calculated the median credit score for each census tract, visualized in Figure 3. The map shows that areas of lower credit scores are in North Minneapolis, parts of Brooklyn Center and Brooklyn Park, and much of St. Paul, whereas suburban neighborhoods that are farther from the central cities tend to have higher credit scores.5 Comparing Figure 3 with Figure 1, we see that neighborhoods with lower median credit scores tend to overlap with neighborhoods where a higher concentration of residents of color live.
Loading map 3...
As in the case of credit inclusion, we then look directly at the relationship between race/ethnicity and credit scores. Throughout the Twin Cities, tract-level median credit scores are positively associated with the share of White population in those tracts. See Figure 4. This means that neighborhoods with a higher share of White population tend to have higher credit scores on average. In neighborhoods with a 75 percent or higher White share of the population, the median credit score averages 787 points, whereas in neighborhoods with a 25 percent or lower White share of the population, the median credit score falls to an average of 679 points.
Loading chart 4...
However, differences in income could account for the variation in credit score across neighborhoods. While income is not considered in the credit scoring system, it could signal an individual’s ability to pay their debts on time, which directly impacts their credit score.
In Figure 5, the blue curve shows the predicted median credit score based on median income alone, with each dot or other plotted shape representing a census tract. Neighborhoods with larger shares of residents of color tend to have both lower median income and lower median credit scores. Furthermore, they tend to be below the curve, suggesting their observed credit score is lower than their predicted credit score based only on their tract’s median household income. In contrast, tracts where a higher share of the population is White tend to have higher median credit scores than would be predicted on the basis of their median incomes. Holding median household income constant, a 10 percentage point increase in a tract’s White population share is associated with a 10-point increase in the tract’s median credit score. (For comparison, without adjusting for income, a 10 percentage point increase in a tract’s share of White population is associated with a 15-point increase in the tract’s median credit score.)
Loading chart 5...
As policymakers and practitioners work to make affordable credit more available to everyone in the Twin Cities area, it is essential to understand exactly where in the region credit gaps are largest. It is also important to understand that credit availability is distributed in ways that mimic other racial disparities in economic conditions, and could potentially reinforce those disparities through patterns of homeownership and other wealth-building activities.
Appendix: About the data
The FRBNY Consumer Credit Panel/Equifax (CCP) data are a nationally representative 5 percent sample of all individuals who have both a Social Security Number and a credit report. All information is anonymized. For more about the CCP data, see the Federal Reserve Bank of New York Staff Report, An Introduction to the New York Fed Consumer Credit Panel.
The credit-inclusion rate estimate requires adult population estimates, which we pulled from the 2020 Decennial Census. We also pulled the census-tract-level race/ethnicity from the 2020 Decennial Census and income information from the 2016–2020 American Community Survey (ACS) to study their relationship with CCP’s credit outcomes.
The 2020 Decennial Census and the 2016–2020 ACS data, however, are reported based on 2020 census tract boundaries that do not perfectly align with the 2010 census tract boundaries included in the CCP data. To merge the two datasets, we first created a one-to-one mapping of 2010 census block boundaries to 2020 census block boundaries using majority rule on the overlapping area between the two versions of the boundaries. (See the IPUMS National Historical Geographic Information System for more information on the 2010 to 2020 census block mapping.) In other words, if one 2010 census block overlaps with two 2020 census blocks with a 30 percent overlap on one and a 70 percent overlap on the other, this particular 2010 census block would be mapped to the 2020 block that has the 70 percent overlap. Since census blocks by definition perfectly align with census tracts, we then aggregate each block up to its corresponding tract. This process gave us the 2010-blocks-to-2020-census-tract mapping. We then merged the CCP data to the Decennial Census and ACS data using this mapping. All analyses were done at the 2020 census tract level.