H.R. 1—the omnibus electoral reform package—is scheduled for a House vote later this week. The managers’ amendment to the bill was just released, and includes (among other things) some substantial (and beneficial) changes to the bill’s prohibition of partisan gerrymandering.

First, the amendment bars intentional (not just excessive) gerrymandering: any congressional plan “drawn with the intent . . . of unduly favoring or disfavoring any political party.” To assist in the determination of partisan intent, the amendment creates a rebuttable presumption that a map that wins significant bipartisan support doesn’t have a partisan purpose.

Second, the amendment clarifies that the severity of a plan’s partisan bias is relevant to the assessment of partisan effect. The bill had previously referred only to the durability of a map’s partisan skew. Severity and durability are distinct concepts, of course, and both are important aspects of an effective gerrymander.

Third, and maybe most importantly, the amendment identifies one way in which a plan can have the effect of unduly favoring a party. Based on quantitative measures of partisan fairness like the efficiency gap, the declination, and so on, a plan must be likely to have a partisan bias exceeding one seat (in states with twenty or fewer congressional districts) or two seats (in states with more than twenty congressional districts). Based on alternative maps (most likely generated by a redistricting algorithm), it must also be possible to design a less biased plan that complies with all nonpartisan legal criteria.

Note that this approach to demonstrating partisan effect doesn’t preclude other ways of establishing liability. Note also the convergence between this approach and the plaintiffs’ proposals in Rucho and other partisan gerrymandering cases. Those proposals generally required a large and durable partisan effect that couldn’t be justified by legitimate factors. That’s what this approach calls for, too, only in somewhat more specific language.