A common claim on both the right and the far left is that Democrats are the party of the rich. How true is this? Well, here’s a plot of US counties by median household income in 2019 and 2020 Democratic presidential margin:

The coefficient of the linear regression is 0.00000632, meaning a $1,000 increase in household income is associated with a predicted 0.63-point INCREASE in Democratic margin. I also plotted 1988 presidential election results on 1990 median household income:

For the 1988 election, the regression coefficient is -0.00000409, meaning a $1,000 increase in household income was associated with a 0.41-point DECREASE in Democratic margin. In 1988, the poorer a county was, the more Democratic you’d expect it to be. Today, it’s the opposite.

When people call Democrats “the party of the rich” they’re generally being a bit too eager with the characterization, but it’s hard to deny the coalitional changes that have occurred over the past few decades, one of which is that richer areas are now, generally, more Democratic.

Already, I know some of you are typing “but education is the real predictor, not income.” This is mostly true; when I ran a multivariate regression on 2020 results (which I may post later), education had by far the highest coefficient of any explanatory variable.

However, even after controlling for education & race, the coefficient of income increased between 1988 and 2020. And regardless of what the “true” cause is, it doesn’t change the reality that Dems increasingly rely on the support of rich areas as they lose support in poorer ones.

Thanks to @thomasfrank_ for getting me interested in this issue with his great book What’s the Matter with Kansas? and @Millenarian22 for inspiring me to do this with his cool regressions. @jmilescoleman @davidshor @kilometerbryman @SenhorRaposa