Cook County's 2026 property tax reassessment on the South Side arrives after a cycle in which the county's machine-learning tax model made assessments fairer by one measure, yet homeowner bills soared. The appeals system, not the algorithm, is the culprit, according to a Forbes analysis of Cook County data [S1]. For South Side homeowners in neighborhoods like Bronzeville, the practical takeaway is that a fair valuation is only the opening offer; what a homeowner finally pays depends on who shows up to argue [S1].

The county's machine-learning model, built under former assessor Fritz Kaegi, estimated what a house would fetch on the open market and taxed it off that number [S1]. A 2025 evaluation by the University of Chicago's Center for Municipal Finance concluded that Kaegi's office had "dramatically reduced" the old regressivity that had overtaxed modest homes, and residential assessments sat inside the fairness range that the assessing profession sets for itself [S1]. Berry, a University of Chicago professor, estimated that the shift saved homeowners roughly $1.9 billion compared with what they would have paid under the old regime [S1].

Yet the model passed its own exam and the assessor who built it was shown the door. In the March 17 Democratic primary, Kaegi lost to challenger Pat Hynes, 52.5 percent to 47.5 percent, with no Republican in the race, meaning Hynes is all but certain to take the office [S1]. The thing that did Kaegi in was the very thing his model was supposed to cure: tax bills that kept climbing for the people who could least afford them [S1]. In West Garfield Park, the typical homeowner's property tax bill had just more than doubled in a single cycle, climbing 133 percent to $3,448, according to an analysis from the Cook County treasurer's office [S1]. A few miles north, in leafy Lincoln Park, the median bill nudged up by less than 6 percent [S1].