The federal Paycheck Protection Program, one of the largest financial bailouts since the Great Depression, promised to provide, in the words of then-President Donald Trump, “unprecedented support to small businesses.” Since its launch in April 2020, the program has injected more than $770 billion into the nation’s businesses, including Reveal from The Center for Investigative Reporting.

Through the Coronavirus Aid, Relief, and Economic Security Act, or CARES Act, Congress ordered the Small Business Administration and the Treasury Department to issue guidance to lenders to ensure that the program prioritized underserved markets. 

Yet a Reveal analysis of more than 5 million PPP loans issued in the program’s first two rounds, in 2020, found widespread racial disparities in how those loans were distributed.

After successfully suing the federal government, with 10 other news organizations, for detailed PPP data, Reveal calculated the proportion of businesses in every census tract that received PPP loans in 2020. In the map below, each tract is coded according to its majority racial group. A dark shade indicates a high loan rate; businesses in pale areas were largely left out by the massive forgivable loan program.


Once the Small Business Administration released detailed data from the Paycheck Protection Program, Reveal began to analyze the first two rounds of the data, covering all program loans for 2020. Reveal asked Geocodio, a commercial geocoding service, to provide geographic coordinates (latitude and longitude) and associated census tracts for those loans. We then aggregated the loan data by census tract for our analysis.

In order to generate PPP loan rates by tract, Reveal had to settle on a way to calculate how many businesses in each tract would have been eligible to apply. In the end, Reveal merged two sources of data to come up with the best estimate of businesses by census tract: (1) a commonly used count of business addresses from the U.S. Department of Housing and Urban Development and the U.S. Postal Service and (2) a count of self-employed workers from the Census Bureau’s American Community Survey.  

The Postal Service provides HUD with quarterly data on residential and business addresses aggregated to the census tract level. We used data from the fourth quarter of 2019 to reflect business addresses before the COVID-19 pandemic began to affect economic activity.

According to Small Business Administration data, sole proprietors, independent contractors and self-employed individuals made up more than one-fifth of PPP loans nationwide last year. Our count of self-employed workers by census tract comes from the American Community Survey 2015-2019.

Both business addresses and self-employed counts include some businesses that would not qualify as small; however, 99.9% of businesses in the United States are considered “small” according to the Small Business Administration, so the impact on our findings would be negligible. 

Using racial demographic data by census tract from the American Community Survey, we classified a racial group as the majority if that group made up more than 50% of the census tract’s population; if a tract had no majority, we labeled the tract “no majority.” Census tracts labeled “N/A” had a population of 25 or less and thus were excluded from our analysis.

Reveal explored other databases, including the Census Bureau’s County Business Patterns and Nonemployer Statistics, to model the count of businesses at a neighborhood level. The disparities remained constant.

This story was produced with the support of the John S. Knight Journalism Fellowships and Big Local News at Stanford University. Download our data and view our full methodology at the Stanford Digital Repository or

This story was edited by Soo Oh and Esther Kaplan and copy edited by Nikki Frick.
Mohamed Al Elew can be reached at

Creative Commons License

Republish our articles for free, online or in print, under a Creative Commons license.

Mohamed Al Elew (he/him) was a data reporter for Reveal. He received his bachelor’s degree in computer science at the University of California San Diego, where he was a research scholar at the Data Science Institute and served as editor in chief of The Triton, the school’s independent student newsroom. As an intern at CalMatters, he worked on an award-winning investigation into instruction lost at California public schools due to natural disasters and infrastructure failures.