[Replication] What the Eighteen-Point Green Backlash Measures: Reproducing and Reframing the Magnitude of Italy's Area B Result

Abstract. Colantone, Di Lonardo, Margalit & Percoco (2024) report that owners of cars banned by Milan's Area B low-emission zone became more likely to vote Lega in 2019—13.5 points on average, an 18.3-point column-2 benchmark—driven by pocketbook loss, not anti-environmentalism. This replication reproduces every published coefficient exactly. The 18.3-point headline is the diesel×Euro-4 interaction, a difference-in-differences in levels collinear with banned-group membership; the simple banned-versus-unaffected contrast is +2 points unconditional (p=.55) and +8 with controls (p=.041), so the headline is roughly 2.3 times a conventional treatment-control difference, significant only once an imbalanced treated group is adjusted. Across the four-party outcome set it softens to a family-wise-corrected p of .081–.088, holding at p=.020 only as the single pre-specified Lega outcome. The mechanism is robust: anti-environmentalism is refuted, the cost dose-response is monotone, and an independent zero-context rebuild predicted +6 points. The finding is robust with scope.