This column by Eric Levitz has been generating some buzz. Based on the congressional districts redrawn so far, the piece argues that “the new House map will be much less biased in the GOP’s favor than the old one.” The new map might even be “tilted, ever so slightly, in the Democrats’ favor.” The “assumption that’s animated the progressive movement’s push for a package of democracy reforms”—the assumption of a large pro-Republican skew due to gerrymandering—is therefore wrong.

I want to push back on both Levitz’s factual claims and their supposed implications. It’s too soon to say how the new House map will compare to the old one. With respect to the districts redrawn so far, they’re only slightly different—not much less biased—than their predecessors. An exclusive focus on national partisan fairness overlooks other redistricting harms. And the case for ending gerrymandering isn’t undercut by the prospect of a slightly less skewed House map in 2022.

1. Prematurity. First off, it’s too early in the redistricting cycle to be reaching firm conclusions about the new House map. Of the forty-four states with more than one congressional seat, twenty-one haven’t yet redrawn their districts. Between them, these states have 194 congressional seats—nearly half the membership of the House. Imminent Republican gerrymanders in Florida, Georgia, and elsewhere will soon push the new House map to the right. A big Democratic gerrymander in New York, if it comes to pass, will have the opposite effect.

2. Continuity, Not Change. Next, it’s simply incorrect that the House districts redrawn so far are “much less biased” than the districts they replaced. The twenty-three states that have completed redistricting have 235 House seats among them. According to PlanScore’s predictive model, in an electoral environment like that of 2020, Democrats would be expected to win 120 of these 235 seats, or 51%. Prior to redistricting, these same twenty-three states had 233 House seats. (California, Illinois, Ohio, and West Virginia each lost a seat; Colorado, Montana, North Carolina, and Oregon each gained a seat; and Texas gained two.) Based on PlanScore’s model, Democrats would likely have won 121 of these 233 old districts, or 52%. Democrats’ expected share of these twenty-three states’ districts has thus dropped by almost a percentage point. Under the new House map, Democrats will probably win one seat fewer in these states than they would have under the old map.

Instead of directly counting Democratic and Republican seats, Levitz focuses on the numbers of districts to the left and right of the national average. According to PlanScore’s model, the mean redrawn district is about 52% Democratic. Of the 235 new districts, 110 are more Democratic than the mean district and 125 are more Republican. So it’s Republicans, not Democrats, who are substantially advantaged relative to the national average. Moreover, the Republican edge in the upcoming cycle (125 to 110) is almost identical to the one they enjoyed prior to redistricting (127 to 106).

To be clear, the numbers Levitz cites aren’t wrong. They just assume that the House vote would be exactly the same as the Biden-Trump presidential vote. However, we know that elections at these two different levels, although highly related, aren’t identical. In particular, Democratic House candidates in 2020 typically ran one or two percentage points behind Biden. PlanScore’s model incorporates this information, as well as potential fluctuations in the parties’ performances, and is therefore more accurate than the raw presidential vote.

3. Other Redistricting Harms. But let’s say Levitz’s figures are right. That is, let’s say the House districts redrawn so far will be perfectly balanced—not (as is actually the case) modestly biased in favor of Republicans. Perfect balance at the national level hardly means that all redistricting problems have been solved. One of these continuing concerns is lack of competition. Observers have been bemoaning the prevalence of uncompetitive House districts for years. In the redrawn House districts, this situation gets even worse. The number of districts expected to be won by fewer than ten percentage points (a traditional definition of a swing seat) drops from 56 to 51. The number of safe districts likely to be won by more than twenty percentage points rises from 122 to 130. The (arguable) presence of national partisan fairness thus does nothing to stop competition from being squeezed out of the system.

Minority representation is another issue unrelated to the House’s overall partisan balance. Unleashed by Shelby County, Republican line-drawers have been cracking and packing minority voters with abandon in previously covered southern states. Cracking has been the dominant Republican strategy in states like North Carolina and Texas: weakening existing minority opportunity districts and refusing to draw new ones despite dramatic growth in the size of the minority population. In states like Alabama, on the other hand, packing has been the weapon of choice: concentrating minority voters in a single district in order to reduce their influence elsewhere. Again, the aggregate national picture obscures these troubling subnational developments.

National partisan fairness is perfectly compatible, too, with extreme subnational partisan unfairness. Partisan balance at the national level can result from severe Democratic and Republican gerrymanders in different states offsetting one another. And in fact, that’s exactly what’s going on. Republicans have gerrymandered states like Alabama, Indiana, North Carolina, Ohio, and Texas. Democrats have gerrymandered states like Illinois, Maryland, Nevada, New Mexico, and Oregon. These gerrymanders still matter—they still cause huge biases in the makeups of particular congressional delegations—even if they add up to a reasonably representative House map.

4. The Case for Reform. The implication of this discussion is that the case for congressional action to fight gerrymandering remains as strong as ever. Contrary to Levitz’s analysis, the redrawn House districts are about as biased in a Republican direction as their predecessors. The redrawn districts also suppress competition, undermine minority representation, and include many glaring gerrymanders that just happen to (mostly) cancel out. And fundamentally, gerrymandering is wrong and should be ended even if it probably won’t be dispositive in the next congressional election. The fate of legislation as sweeping as the Freedom to Vote Act shouldn’t hinge on something as narrow as the bill’s likely effects in 2022.