Where the money and the vote disagree
Which counties vote one way while their reliable donors lean the other — and who lives there?
Each county’s 2020 presidential vote, and the lean of its reliable individual donors. Hue is the vote; saturated counties are those whose donors lean the other way. Brush the scatter to find them on the map.
Across the country, a county's donors usually lean the way it votes (population-weighted r = −0.85). In 20% of counties they lean the other way, and those counties cluster. Select a cell of the key, a preset, or drag across the scatter.
2020 vote →
County data · 2,839 rows
| County | FIPS | Dem share (%) | Donor CFscore | Cell | Black (%) | Reliable donors | Population |
|---|---|---|---|---|---|---|---|
| County 26145 | 26145 | 50.1 | -0.03 | Voted Democratic, donors lean left | 18.2 | 2,611 | 190,124 |
| County 26065 | 26065 | 66.4 | -0.33 | Voted Democratic, donors lean left | 11.1 | 9,117 | 284,900 |
| County 26139 | 26139 | 39.1 | 0.07 | Voted Republican, donors lean right | 1.5 | 3,634 | 296,200 |
| County 26125 | 26125 | 57.1 | -0.22 | Voted Democratic, donors lean left | 12.9 | 33,789 | 1,274,395 |
| County 26005 | 26005 | 37.1 | -0.07 | Voted Republican, donors lean left | 1.2 | 1,465 | 120,502 |
| County 06075 | 06075 | 87.0 | -0.87 | Voted Democratic, donors lean left | 4.8 | 45,868 | 873,965 |
| County 12073 | 12073 | 64.3 | -0.19 | Voted Democratic, donors lean left | 30.5 | 8,772 | 292,198 |
| County 17085 | 17085 | 41.6 | -0.41 | Voted Republican, donors lean left | 0.7 | 267 | 22,035 |
| County 26093 | 26093 | 38.5 | -0.04 | Voted Republican, donors lean left | 0.6 | 3,495 | 193,866 |
| County 25017 | 25017 | 73.1 | -0.64 | Voted Democratic, donors lean left | 4.8 | 65,745 | 1,632,002 |
| County 25023 | 25023 | 58.9 | -0.28 | Voted Democratic, donors lean left | 8.4 | 11,536 | 530,819 |
| County 29189 | 29189 | 62.2 | -0.27 | Voted Democratic, donors lean left | 24.1 | 18,470 | 1,004,125 |
| County 55031 | 55031 | 54.8 | -0.23 | Voted Democratic, donors lean left | 1 | 709 | 44,295 |
| County 37081 | 37081 | 61.7 | -0.2 | Voted Democratic, donors lean left | 33.7 | 6,703 | 541,299 |
| County 26037 | 26037 | 46.7 | -0.05 | Voted Republican, donors lean left | 1.8 | 1,513 | 79,128 |
| County 36061 | 36061 | 87.6 | -0.72 | Voted Democratic, donors lean left | 12.4 | 82,833 | 1,694,251 |
| County 24023 | 24023 | 21.5 | 0.14 | Voted Republican, donors lean right | 0.9 | 287 | 28,806 |
| County 42003 | 42003 | 60.3 | -0.36 | Voted Democratic, donors lean left | 12.6 | 22,455 | 1,250,578 |
| County 36047 | 36047 | 77.6 | -0.82 | Voted Democratic, donors lean left | 28 | 53,300 | 2,736,074 |
| County 06013 | 06013 | 73.1 | -0.71 | Voted Democratic, donors lean left | 8.1 | 32,404 | 1,165,927 |
Donor lean is the mean CFscore of a county's reliable individual donors in DIME v4.0 — people who gave to at least eight distinct recipients, located at their most recent address; negative is left. Counties with fewer than 25 reliable donors, and Alaska, have no estimate and stay grey. The rules at 50% and 0 are the class borders, so a dot's quadrant is its county's colour. Because the national donor pool leans left, a Republican county's “left” donors are often only just left of zero; the firmer divergence is the Democratic counties with donors on the right. Among Democratic-voting counties, donor lean and Black share correlate at +0.63. Adding donor lean to a demographic model of the vote raises R² from 0.60 to 0.79 — an association, not a cause. Strictly descriptive: a county's donor mean says nothing about any one resident.
Stanford DIME v4.0 (reliable individual donors, county means) × 2020 county presidential returns · 2,839 counties
Reading notes
Strictly descriptive county aggregates: a donor mean says nothing about any one resident. Counties with fewer than 25 reliable donors, and Alaska, are suppressed. The national donor pool leans left, so a Republican county’s “left” donors are often only just left of zero.
Source & method
Joined the chapter's `map` rows (FIPS, cell, donor CFscore, Democratic two-party share, % Black, reliable-donor count) to its `scatter` rows (population) by position, after asserting on every row that scatter.x equals dem2p, scatter.y equals the donor CFscore, and % Black agrees — the scatter rows carry no FIPS. Kept all 2,839 published counties (the source suppresses counties with fewer than 25 reliable donors and excludes Alaska; those stay missing, never zero). Kept the two named lists (12 + 12 counties with at least 75 reliable donors), the horse-race R² summary, both correlations and the divergent share. Dropped the duplicated `annotations` block. Values are as published (three decimals for shares and CFscores, one for % Black); nothing is re-estimated. The cell classification is the source's, and is checked against the sign rule it is described by.
Capture provenance
- box:Politics / analysis/the-donor-map-vs-the-voter-map/data.json
- Modified 2026-06-27 · SHA-256 c33b38bef55c1a5d34b9d1827d9175446ae9fa3949f8dada4e1c88ec206cf8b5
- Captured
- 2026-09-12
- Fixture SHA-256
- b228c4234bfdc5251d8a2514d1752713a6eeb20e85a8e80f50bc2fc6dbde8461
- Reuse note
- Donor ideology: Bonica, Adam. Database on Ideology, Money in Politics, and Elections (DIME), v4.0, Stanford University — county aggregates of reliable individual donors only. Vote: 2020 county presidential returns. Cite DIME (Bonica 2014, AJPS 58(2)) when reusing.
