About

The Home-mortgage Loan Denial Gap measures the disparity in the rates at which mortgage loan applications are denied for Black individuals compared to White individuals in a county. This metric includes data from both Hispanic and non-Hispanic ethnicities. According to data from the Home Mortgage Disclosure Act (HMDA) in 2022, Black applicants experienced a denial rate of approximately 16%, while White applicants had a denial rate of around 7%. This indicates that Black applicants are more than twice as likely to be denied a mortgage loan as White applicants, highlighting a significant disparity in access to homeownership ¹.

Why is the Home-mortgage Loan Denial Gap important to the Structural Racism and Discrimination (SRD) Index?

Persistent racial disparities in home-mortgage loan denial rates in the United States are empirically linked to both historical structural racism – such as redlining – and ongoing institutional discrimination, with robust evidence showing that these gaps cannot be fully explained by applicant financial characteristics alone ^2-7. Studies demonstrated that neighborhoods historically redlined or marginalized by the Home Owners’ Loan Corporation (HOLC) maps in the 1930s-40s show significantly higher denial rates for mortgage loans today, even after controlling for neighborhood disadvantage and applicant characteristics ^2,4. For instance, research suggests that a majority of observed denial-rate differences between Black, Hispanic, and White applicants remain unexplained by observable factors such as income, loan-to-value, and debt-to-income ratios, indicating the presence of systematic discrimination beyond individual credit risks ^6,7. This measure is crucial for understanding structural barriers to homeownership, which directly impact long-term wealth accumulation and neighborhood stability. Addressing these inequities is essential to reduce racial disparities in housing access and ensure fair financial opportunities for all residents.

What is the expected relation to Structural Racism and Discrimination?

A higher value of the Home-Mortgage Loan Denial Gap between Black and White applicants contribute to the higher value or score of the SRD Index.

How is the Home-mortgage Loan Denial Gap calculated?

Data Source

We obtained data from the Federal Financial Institutions Examination Council (FEIEC) ⁸ and  Home Mortgage Disclosure Act (HMDA) ⁸. The data is publicly available.

Data

The HMDA Loan Application Register (LAR) file contains data on mortgage loan applications, including details about the applicants’ demographics, actions taken, and geographic information. Using individual applicant data, the information was aggregated at the county level for Black and White applicants based on the outcome of loans that were not originated. We used the following four variables at the county level.

Variables*	Year	Unit
Total Black or African American Loan Applicants	1990 | 2000 | 2010 | 2020	Number
Total White Loan Applicants	1990 | 2000 | 2010 | 2020	Number
Number of Black or African American applicants whose loan was not originated	1990 | 2000 | 2010 | 2020	Number
Number of White applicants whose loan was not originated	1990 | 2000 | 2010 | 2020	Number

* Individuals from both Hispanic and non-Hispanic ethnicities are included.

Methodology

We calculated the Home-Mortgage Loan Denial Gap using a ratio formula:
$$
RPrBlWhLNO = \frac{PrBlLNO}{PrWhLNO}
$$

Where:
RPrBlWhLNO: Ratio of proportions for Black to White applicants whose loan was not originated
PrBlLNO = BlLNO/BlAp: Proportion of Black or African American applicants whose loan did not originate among Black loan applicants.
PrWhLNO = WhLNO/WhAp: Proportion of White applicants whose loan did not originate among White loan applicants.

Missing Data

We replaced missing values (where applicable) by using functions that were based on Tobler’s First Law of Geography, i.e., “everything is related to everything else, but near things are more related than distant things” ⁹. This law also informs the concept of Spatial Autocorrelation (SA). The variables were moderately spatially autocorrelated and the missing value in a county was imputed using the median value of the adjacent neighbors. The adjacent neighbors were identified using the ‘PolygonNeighbors’ tool in Python’s arcpy.analysis module ¹⁰. The adjacency is defined by any common boundary or vertex between two polygons. The following table shows a summary of the final missing data.

Year	Number of Counties with no data
	Black Loan Denial	White Loan Denial
2020	9	9
2010	0	0
2000	1	0
1990	12	11

Limitations

For 2020 and 2010, loan non-origination or loan denial includes application denied, file closed for incompleteness, and preapproval request denied. For 2000 and 1990, loan non-origination included application denied and file closed for incompleteness.

References

1. Consumer Financial Protection Bureau. (2023). Home Mortgage Disclosure Act (HMDA) data.

2. Faber, J. (2021). Contemporary echoes of segregationist policy: Spatial marking and the persistence of inequality. Urban Studies, 58(5), 1067-1086.

3. Holloway, S. R., & Wyly, E. K. (2001). ” The Color of Money” Expanded: Geographically Contingent Mortgage Lending in Atlanta. Journal of Housing Research, 55-90.

4. Namin, S., Zhou, Y., Xu, W., McGinley, E., Jankowski, C., Laud, P., & Beyer, K. (2022). Persistence of mortgage lending bias in the United States: 80 years after the Home Owners’ Loan Corporation security maps. Journal of race, ethnicity and the city, 3(1), 70-94.

5. Holloway, S. R. (1998). Exploring the neighborhood contingency of race discrimination in mortgage lending in Columbus, Ohio. Annals of the Association of American Geographers, 88(2), 252-276.

6. Assad, J. C. (2016). The Role of Income and location in Racial/Ethnic differences on loan denial in three Mississippi Counties. Journal of Economics and Development Studies.

7. Park, K. A. (2021). Measuring risk and access to mortgage credit with new disclosure data. Journal of Structured Finance, 26(4), 53-72.

8. Federal Financial Institutions Examination Council (FFIEC). (n.d.). Home Mortgage Disclosure Act (HMDA) data.

9. Tobler, W. R. (1970). A computer movie simulating urban growth in the Detroit region. Economic geography, 46(sup1), 234-240.

10. Esri. (n.d.). Polygon neighbors (analysis). ArcGIS Pro Tool Reference. https://pro.arcgis.com/en/pro-app/latest/tool-reference/analysis/polygon-neighbors.htm