Dem
47
GOP
53
Gerrymandering goes back to the days of Gov. Elbridge Gerry and his salamander-shaped legislative districts. It is an American tradition, although England's "rotten boroughs" might have provided an inspiration. It hasn't always worked out as planned.
Consider the 1894 election as a case study in what can go wrong. In Gerrymandering 101, you learn how to spread your voters over more districts, giving your side a small (or small-ish) majority in many districts instead of a large majority in fewer districts. The hard part is figuring out how small a majority you can get away with and still win. Sometimes politicians get too greedy and inadvertently create a dummymander.
Back in the 19th century, states changed their maps all the time, not just after the census. Between 1862 and 1896, in only one year were all the maps the same as in the previous cycle. Whenever a party took power in a state, it drew new maps. For example, in 1872, Republicans flipped 64 House seats, only to lose 94 seats in 1874.
Gerrymandered districts tend to be competitive. Why draw a district where 60% of the voters are on your side? Why not make it 55% or 53% and use the excess voters to help win some other district? In the late 19th century, something like 40% of House races were won by 5 points or less. After the 1890 census, Democrats didn't shore up their own incumbents. Instead they went for broke.
The year 1894 bears a striking resemblance to the current situation, except with the parties flipped. The Democrats had a president, Stephen Grover Cleveland, who had just won a second nonconsecutive term, tariffs were a hot issue, and the president sent federal troops to Chicago to quell a railroad strike (over the objections of the governor of Illinois). The economy was a big issue then as now.
The maps the Democrats had ever-so-carefully drawn after the 1890 census turned what would have been a mild setback into a bloodbath. For example, the map of Missouri gave the Democrats 13 of the 15 House seats in 1892. When their vote share dropped 6 points in 1894, they lost 8 of those seats. In New York, they went from 20 to 5 seats. Throughout the Northeast, they went from 44 House seats to 7. In the Midwest it was worse. They went from 44 seats to 4 seats. The Democrats lost everywhere outside the South. Nationally, the Democrats lost 114 of the 357 House seats (32%).
There was a depression in 1893, so Cleveland's party was going to be whacked no matter what, but the carefully drawn 1890 maps made it much worse. One calculation based on a more neutral map after the 1890 census showed that Democrats could have held their losses to 59 seats if they hadn't been so greedy. They simply made their margins too thin in order to get a small edge in multiple districts.
In the 20th century, things changed. Many states added provisions to their Constitutions banning mid-decade redistricting. Some banned all redistricting, even after a new census, which led to districts with huge population imbalances as cities grew in population and rural areas shrank. These laws were designed to keep political power in rural areas, even as they were no longer the majority.
Now we seem to be back to 1894, and there could be dummymanders in some states. If Democrats win big next year, people will look back and see two causes. First, the models were wrong. What redistricting apps do is use the precinct-level data from the previous House election to allow the manderer to move precincts in and out of districts to get the desired percentage. The implicit assumption is that precincts don't change their spots between elections. But it is already clear that many minority voters who supported Trump in 2024 are either going to support the Democrats in 2026 or not vote at all. Either way, giving yourself a 4-point margin in some district and then having 5% of your voters stay home could spell disaster.
Second, Democrats could be much more energized this time than last time. Few Democrats truly loved Kamala Harris in 2024, but many Democrats truly hate Donald Trump now. This could increase Democratic turnout over a normal midterm year. Combining aggressive midterm gerrymandering, bad models, energized Democrats, and a poor economy, 2026 could be as bad for Republicans as 1894 was for Democrats. OK, the GOP is probably not going to lose 114 seats, since there just aren't that many in-play districts (even if you assume a blue tsunami). But they could still suffer a huge loss, enough to put the House Republican Conference in the wilderness for several cycles (or more). (V)