It’s been widely claimed that Zohran Mamdani (1) disproportionately drew votes from affluent and educated professionals while doing poorly with working-class voters, and (2) that he severely underperformed among blacks. The second claim is true — and also helps explain why the first claim is false.

If you split New York City’s 2,168 Census tracts into two equally populated halves, based on their black population shares — a “blacker” half and a “less black” half — the pattern of class support for Mamdani turns out to be almost exactly the opposite in the two halves.

In the “less black” half, the Mamdani vote was strongly, consistently, positively correlated with working-class status. He did best with janitors and home health aides and worst with lawyers and finance people. The opposite was the case in the blacker half: within those areas Mamdani performed best in the more affluent neighborhoods and worst in the poor areas. See the charts below: