This report includes data from a variety of complex and non-traditional data sources, each with their own limitations and caveats. This appendix is provided as a resource for describing margins of error and other limitations and measures of uncertainty inherent in any dataset obtained through survey or derived through statistical calculation.

While all datasets used in the report have been vetted for accuracy and represent the best-known estimates available for the issue referenced, readers should take care to note where margins of error prevent a precise reading or interpretation of data. Following the description of limitations and caveats for each data source, this appendix provides guidance on which tables and figures utilize the data. It should be noted that additional data points mentioned throughout the report may also be related to the aforementioned sources, even if not explicitly identified here.

All comparisons presented are based on direct estimates, even when derived from surveys with associated margins of error. Users should exercise caution when making definitive comparisons between data points that might have overlapping margins of error and can use The Data Center’s Margin of Error Calculator to determine if two data points are meaningfully different, given stated margins of error. To see the data behind this report, including associated margins of error and other related data, please explore the accompanying data download [link to spreadsheet]. For questions about the data sources or methodology, contact Kindred Futures (belovedresearch@kinfutures.org) or The Data Center (questions@datacenterresearch.org).

State and Local Wealth Estimates

Disaggregated data on wealth is provided by The Data Center, based on the 2018 Survey of Income and Program Participation (SIPP) and the 2018 American Community Survey Public Use Microdata Sample (ACS-PUMS). Given the limited disaggregated wealth data available at geographies smaller than the national level, this data is derived from modeled estimates based on 2018 data designed to align with the most recent complete SIPP survey that does not include data quality warnings.

This report breaks down wealth estimates by metropolitan status, which is used as a proxy for rurality. The definition of metropolitan areas aligns with the USDA Rural-Urban Continuum Codes. While this definition may not fully capture the continuous and nuanced nature of rurality, particularly in areas adjacent to urban areas, it offers a consistent classification of census geographies to allow for comparisons of wealth differences.

The Data Center’s wealth estimates include margins of error, which are provided in the accompanying data file to this report. Users should be aware of data points with overlapping margins of error when making definitive comparisons between estimates, as such differences are unlikely to be statistically significant.

Experienced users of wealth data might note that The Data Center’s modeled wealth estimates may differ from other published sources. These differences are due to methodological choices and underlying data sources. In particular, while SIPP is designed to measure wealth in detail, it includes a smaller sample than the ACS and lacks geographic specificity. The Data Center has verified that differences between modeled estimates and state-level SIPP data are due to compositional differences in the surveyed population between the SIPP and the ACS, with the ACS being a more comprehensive survey of the sub-national population. For more information regarding the methodology used to produce these estimates, see the technical paper accompanying this data.

Survey of Income and Program Participation

The Survey of Income and Program Participation (SIPP) is a nationally representative longitudinal household survey administered by the Census Bureau which reports income, employment, financial status, household characteristics, and wealth. While SIPP data is available at the state level, estimates become less reliable as they are disaggregated due to small subgroup sample sizes. Users should keep this in mind when using SIPP estimates and consider the accompanying margins of error when making comparisons.

Financial Health and Wealth Dashboard

The Urban Institute’s Financial Health and Wealth Dashboard compiles multiple data sources to present disaggregated measures of household financial health, including liquid assets, debt, and net worth and highlights racial disparities in financial well-being. The dashboard relies on statistical models to generate net worth estimates and emergency savings, similar to other sources that utilize statistical models when there is limited data availability on assets and debts at more granular levels. This report utilizes the dashboard’s state-level data on delinquent debt and emergency savings.

While the dashboard’s methodology is similar to other modeled estimates of assets and debts, small differences in methodology can result in different overall values. For example, The Urban Institute provides wealth data at the city level; but with increased granularity comes decreased sample sizes across various demographic groups. This tradeoff results in city-level estimates disaggregated by race only being reported for communities where that race constitutes the majority. This is an effective way to avoid sample size problems for specific demographics, but likely over-emphasizes the racial skew typically inherent in wealth accumulation. The Urban Institute’s variety of estimates in the Financial Health and Wealth Dashboard provides helpful context about financial characteristics in communities with distinct demographic profiles.

American Community Survey

The American Community Survey (ACS), a national survey conducted by the U.S. Census Bureau, provides annual estimates on demographic, social, economic, and housing characteristics. ACS data is available at various geographic levels, though estimates for smaller geographies or subgroups can include substantial margins of error. This is generally not of concern when data is presented at the state level due to larger sample sizes, but users should be cautious when comparing subgroups or small populations over time or across geographies. Data points with margins of error can be found in the accompanying data download file.

Decennial Census

The Decennial Census is a full count of the U.S. population conducted every ten years by the U.S. Census Bureau. It provides data on total population, housing units, and demographic characteristics. While limited in scope compared to the ACS, the Decennial Census is considered a complete count of the population and thus does not have associated margins of error.

Annual Business Survey

The Annual Business Survey (ABS) is a U.S. Census Bureau product which provides detailed data on the demographic characteristics of business owners by race, ethnicity, and sex at various geographies. While the ABS is limited to employer businesses and therefore does not capture the full landscape of informal or non-employer businesses, it is available over time and can give insight into the diversity of business ownership in an area. Note that employer businesses in any metro area will account for the vast majority of business revenues, but only a fraction of the actual businesses in the region.

IPUMS USA

The Integrated Public Use Microdata Series (IPUMS USA) provides anonymized individual- and household-level data from the U.S. Census and American Community Survey (ACS). This data allows for more detailed disaggregation by race, income, age, and geography than the ACS summary tables, but is subject to higher margins of error in smaller subgroups. Users should be cautious when interpreting differences between groups with overlapping margins of error. For example, while Figure 7 shows a higher median home value for White households in North Carolina compared to South Carolina, the overlapping margins of error suggest this difference is unlikely to be statistically significant. The associated margins of error for all IPUMS USA estimates are included in the accompanying downloadable data table.

IPUMS NHGIS

The National Historical Geographic Information System (NHGIS) is a product of IPUMS that harmonizes census data across time and geographies, including Decennial Census variables dating back to the early 20th century. NHGIS provides a consistent framework to examine demographic and housing trends over time. Since most data points included from NHGIS come from Decennial Census, they typically do not have associated margins of error.

Renewing Inequality and Urban Renewal Project Characteristics

The University of Richmond’s Renewing Inequality project compiles information from the federal government’s Urban Renewal Project Characteristics and other various sources from 1955 to 1966 to assess the impact of urban renewal projects on family displacements in select U.S cities.

The authors of the project emphasize that this dataset represents only a fraction of displacements during this period, and that the figures reported to the federal government are often estimates and are considerably lower than figures documented by other urban renewal reports. Additionally, the data reflects displaced families rather than individuals that were displaced through urban renewal.

Several southern cities were selected for inclusion in this report: Atlanta, GA; Birmingham, AL; Jackson, MS; Little Rock, AR; Memphis, TN; Mobile, AL; New Orleans, LA; and Savannah, GA.

Another project by The University of Richmond, Mapping Inequality: Redlining in New Deal America, provides estimates of the percentage of census tracts at the state level that were redlined during the 1930s and 1940s.

Federal Deposit Insurance Corporation

The Federal Deposit Insurance Corporation (FDIC) conducts the biennial Survey of Household Use of Banking and Financial Services, which assesses financial inclusion, such as rates of unbanked and underbanked households at the state level. When the sample size for a state is too small to produce precise estimates, an NA value is presented.

Home Mortgage Disclosure Act

The Home Mortgage Disclosure Act (HMDA) dataset includes detailed records on mortgage applications, approvals, denials, interest rates, and applicant demographics. It should be noted that HMDA does not capture all lenders, and there may be inconsistencies in self-reported demographic information. Data by race is presented for individuals identifying as White alone and Black alone.

Federal Reserve System Banking Deserts Data

This report includes analysis by Kindred Futures of the Federal Reserve System’s Banking Deserts dataset, which identifies census tracts with limited access to brick-and-mortar banking institutions. The dataset is designed to highlight areas where residents may face significant barriers to financial services, such as lack of proximity to bank branches or concentrations of unbanked households. Kindred Futures used this data to examine the geographic distribution of banking deserts across Southern states, particularly in areas with high Black population density. While the data provides important insight into structural gaps in financial infrastructure, users should note that it does not account for informal financial services or access to online banking, which may vary across regions and populations. The original dataset can be accessed through the Fed Communities Banking Desert Dashboard.

National Equity Atlas

The National Equity Atlas is a collaboration between PolicyLink and the USC Equity Research Institute that provides data on racial and economic equity indicators across U.S. geographies. This report utilizes the National Equity Atlas’ indicator on Occupational Segregation, which analyzes 2018-2022 IPUMS USA data to calculate the percentage of workers aged 16+ by race and ethnicity in each occupational group by state. Standard limitations described for the IPUMS USA data apply to this indicator; however, concerns related to sample size are often mitigated by aggregating data across multiple years, as done in this analysis.

Uniform Appraisal Dataset (UAD) Aggregate Statistics

The Uniform Appraisal Dataset (UAD) Aggregate Statistics, released by the Federal Housing Finance Agency (FHFA), provide standardized property-level appraisal data collected through mortgage underwriting processes. These data allow for analysis of appraisal trends across geographic and demographic contexts, including patterns that reflect racial and neighborhood disparities in home valuation. For this report, Kindred Futures analyzed the 2025 UAD Aggregate Statistics to examine appraisal values across communities in the South. While the UAD offers extensive data coverage, it is limited to properties appraised for mortgages purchased or guaranteed by Fannie Mae or Freddie Mac, and does not capture cash sales or loans held in portfolio. Users should note that this dataset is best used to observe aggregate trends rather than individual-level outcomes. More information is available via the Federal Housing Finance Agency.

FEMA National Risk Index

This analysis draws on two publicly available federal data sources. First, the U.S. Census Bureau’s 2023 American Community Survey (ACS) 5-Year Estimates were used to calculate the percentage of the Black population at the census tract level. The ACS provides annually updated demographic, social, economic, and housing data, and is considered reliable for geographic comparisons at the tract level due to its multi-year sampling approach. Second, the Federal Emergency Management Agency’s (FEMA) National Risk Index was used to incorporate each tract’s Community Resilience Score—a composite metric that reflects a community’s capacity to prepare for, respond to, and recover from disasters based on socioeconomic and infrastructure indicators. Kindred Futures combined these datasets to explore how Black population density correlates with resilience scores across communities in the South. Users should note that Community Resilience Scores are modeled and comparative, not absolute measures of resilience.

Housing Wealth Gains Estimates

This analysis uses custom tabulations of the U.S. Census Bureau’s 2018 Survey of Income and Program Participation (SIPP), prepared by The Data Center for Kindred Futures. These tabulations provide state-level estimates of median housing equity by race, enabling the calculation of potential wealth gains if Black homeownership rates matched each group’s share of the state population. To estimate the number of Black households and their population shares, the analysis relies on the U.S. Census Bureau’s 2023 American Community Survey (ACS) 5-Year Estimates. Wealth gains are expressed in 2023 dollars. While these modeled data offer valuable insights into structural housing inequities, users should interpret results with caution. The underlying estimates are sensitive to assumptions about home values, equity shares, and racial population distributions, and are not intended to serve as precise forecasts.