The headline: People in counties that voted for Clinton in 2016—Clinton counties—have higher average salaries, pay more federal income taxes, and received greater federal spending than Trump counties.
The details: Individuals in Clinton counties earned 28% more than in Trump counties ($40,111 vs. $31,238), paid 55% more in federal income taxes ($9,087 vs. $5,859), and received 26% more in federal spending ($12,743 vs. $10,093). Federal income taxes account for only 48% of federal tax revenue, hence the large difference between federal spending data per person and federal taxes (payroll taxes account for 35%, corporate income tax for 9%, and excise, state, and other taxes for the final 8%)1. In terms of predicting the 2016 elections, federal spending, salaries, and federal income taxes enter our Illuminating America series in fifth, sixth, and seventh place respectively, explaining between 27 to 36% of the result. The difference between Clinton and Trump counties has a 100% confidence for all three variables.
The data: Tax data was sourced from the IRS2 and federal spending data from the USDA’s Economic Research Service (ERS)3. Tax data from 2010 was used to ensure tax and spending data was aligned (2010 is the latest year for county-level federal spending data).
The data on spending by the US govt by location seems be suspect to me. It was last published in 2010 and discontinued afterwards. The report itself notes that many US Federal spending cannot be identified by location. Federal expenditures for defense and payments for interest on the national debt, state level block grants come to mind that are not easily identified down to the county level.
Yes, good point. I am sure there are lots of uncertainties and errors! But my guess is that it largely makes sense. I don’t think managing the debt is a category in the database. A fair amount of the DOD spending shows up OK (eg, in Virginia!). When I first coded the data into categories (probably almost a year ago) there were lots of items that were a guess as to which category to put it in… but the large majority were OK (I probably binned it into 15 categories or so for thousands and thousands of entries)
I’m still rooting for chlamydia…
Chlamydia’s days as #1 are numbered!