Reveal used geocoding and Census data appends from Geocodio as part of their analysis of the Paycheck Protection Program
Reveal, the non-profit newsroom from the Center for Investigative Reporting, used the Paycheck Protection Program data geocoded and enrich with Census data by Geocodio for their analysis of where PPP loans were granted.
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.
Geocodio released a public dataset of the PPP data, which Reveal then used for their analysis:
Methodology
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.