THE PROMISE of democracy is a system where leaders represent their voters, and where those voters can throw them out when they don’t. But in America, the reverse can also be true: politicians in many states get to choose which voters represent them. The constitution originally delegated the power of drawing the boundaries of electoral districts to the state legislators; many abuse this power and “gerrymander” districts in their party’s favour. Citizens in some states have amended their laws to give watchdog powers to the judiciary or require plans to be approved by an independent third party. But a majority of Americans still live in states where politicians are incentivised to rig elections by drawing biased maps.
Because Republicans have lately controlled more state governments than Democrats, they have tended to do better in congressional redistricting. After redrawing maps in 2010, the Republicans lost the popular vote for the House of Representatives but still won a majority of its seats. In 2012, Mitt Romney’s vote share in the median House district—based on the vote share for Mr Romney—was 0.3 percentage points higher than Barack Obama’s, even though Mr Obama won the national popular vote by 3.9 points. In a perfectly equal system, those numbers should roughly match. In fact the House was biased against Democrats by four points in 2012, and again in 2016. And in 2020 the median district was biased against Joe Biden by two points.
But in 2022, after states redraw their congressional lines according to new population estimates from the 2020 census, the difference between Mr Biden’s margin in the median district in the new plans and his margin in the national popular vote is almost certain to be much closer to zero. Democrats have, by this measure, made surprising gains in the redistricting process. Three main reasons explain their success.
The first is gerrymandering of their own. Although Republicans have been more notorious for abusing the process over the past decade, Democrats have caught on. To illustrate this, look at the new congressional map in New York, which Democrats have just signed into law. According to analysis by The Economist of election-results data collected by FiveThirtyEight, a data-journalism website, the state’s previous plan contained 18 districts where the average vote share for Democratic presidential candidates over the past two contests was higher than Republicans’ vote shares. The new plan has 22.
The story is similar in New Mexico. Democrats control the state government there, too. Whereas the state’s previous congressional districts included one heavily Republican seat and two Democratic strongholds, lawmakers have rejigged the boundaries so all now lean towards the Democrats. The two revamped seats will be competitive in national races in which Republicans do well, but most of the time Democrats will win all three of the state's districts.
That brings the party’s net gains from redistricting in New York and New Mexico to five Democratic seats. Add to that the changes in Oregon and Illinois—both blue states with minimal third-party oversight over redistricting—and Democrats come out with an expected ten seats more than in 2020.
The Democrats have also benefited from reforms that keep districts fairer. Take Colorado. After the releases of the 2000 and 2010 censuses, state legislators there failed to agree on maps, so state courts drew the lines instead. In November 2018, fed-up Coloradans passed Amendment Y via a ballot initiative that authorised an independent commission of residents—four Democrats, four Republicans and four “unaffiliated”—to redraw the state’s congressional boundaries. It joined Arizona and California (which began using similar models in 2000 and 2010, respectively) and Michigan (which adopted such a system in 2018) as the only states using partisan-balanced commissions of citizens to draw district maps.
Colorado’s commissioners were selected in a semi-random fashion. Panels of retired judges and state legislators select applications in eight stages, which involve lotteries and the hand-picking of impressive candidates. The 12 commissioners began meeting in March, and by the final months they were spending as many as 40 hours each week to meet their October deadline. Simple rules guided them: create non-discriminatory, contiguous districts of equal size; maximise political competition; and preserve so-called communities of interest—neighbourhoods where voters have shared policy concerns and would benefit from being in the same district (examples include ethnic, racial and economic groups).
Once submitted, these commission-drawn maps do not require approval from the governor or state legislature. This year, their final map preserves the partisan leanings of the state’s seven existing districts, and adds one more Republican-leaning seat with the additional representative the state gained from redistricting. Further levelling the playing field, the commission also made the seventh district more competitive. The current representative there, Ed Perlmutter, a Democrat, announced last month he would not seek re-election. In a very red year he may have struggled to win the revised district anyway.
The maps in Michigan ended up similarly balanced. Whereas previous boundaries gave an edge to Republicans in eight of the state’s 14 seats—despite voters there being split nearly 50-50 in presidential elections—the new boundaries give them a solid lead in just six. Republicans have the lead in eight, down from nine due to reapportionment.
Last, there are the courts. The judiciaries in Ohio and North Carolina both struck down maps they said favoured Republicans too much, constituting partisan gerrymanders that violated state constitutions. In Ohio, Republicans drew a map which made 73% of the state’s districts lean towards Republicans—though the party won only 55% of the vote statewide. “By any rational measure,” Justice Michael Donnelly said in the court’s opinion, “that skewed result just does not add up.” The new map in North Carolina was even redder. Republicans there passed a map which gave them the edge in 71% of seats, even though Donald Trump beat Mr Biden by only one point there in 2020. Neither state has yet approved a new map.
A lower court in Alabama has also ruled that its new maps are biased against African-Americans. The state has appealed against the order to the Supreme Court, which could issue its decision at any time. As drawn, the map provides for only one of seven districts where black voters will have a chance to elect a candidate of their choice, in a state where 27% of the voters are black. Plaintiffs say legislators could easily have drawn another. Victory could mean another Democratic district in the state, though it is unclear which way the justices will lean.
Cracking and packing
All three of these factors contribute to Democratic gains. Adding up the tally in the 30 states that have passed new maps so far, Democrats have gained 12 new seats where they make up more than half of a district’s voters in presidential elections. Excluding close districts where neither party has an expected lead of five points over the other party, Democrats are up by 11. There are eight fewer competitive seats across these states.
Yet it is important to consider the baseline. Congressional maps for the past decade have been significantly biased against the Democrats. The party has managed to claw back some seats this year, but it is possible that the map will still be biased towards Republicans; the gap between Mr Biden’s margin nationally and in the new median is likely to be about half a percentage point rather than two points.
Further, a simple tally of seats is also not enough to assess partisan balance fully, some political scientists contest. Nick Stephanopolous, a Harvard professor who helps with analysis at PlanScore, a website run by political scientists and statisticians that checks maps for fairness, prefers to focus instead on a statistic called the “efficiency gap”. This measures the share of Democratic and Republican votes that are “wasted” in each seat. Wasted votes are those that are cast for parties in excess of the margin they need in districts they win, plus every vote cast for the party in seats they lose. The measure can be thought of as quantifying both how many members of one party are “cracked” into losing districts and the number that are “packed” into safer ones.
In the states that Joe Biden won in 2020, there is an efficiency gap of roughly five percentage points favouring Democrats. That means roughly 5% more Republican votes than Democratic ones are wasted across them. That is an increase over the two-point gap in those states under previous lines, reflecting Democratic gains from the gerrymandering, reform and litigation efforts.
But in Republican states, the efficiency gap will be roughly 14 points (the precise number depends on what exactly the final maps look like; our number is based on the average across the likeliest plans under consideration). That is an increase over the 13-point efficiency gap in those states under the 2020 congressional map. In other words, the maps in both Democratic and Republican states have become more biased as a result of this decade’s redistricting. And the average Republican plan creates many more wasted votes than the average Democratic one. The only good news was in states with independent redistricting commissions. There, the absolute efficiency gap fell by nearly three points, meaning districts were less gerrymandered. It increased by three points elsewhere.
According to estimates of district partisanship produced by Mr Stephanopolous and shared with The Economist, redder states have also focused more on ring-fencing their current advantage. That has meant shifting seats with mild Republican advantages farther right, instead of focusing on drawing Democrats out of the ones they have. In Texas, for example, the Republican-dominated state legislature packed Democrats into five new seats, but simultaneously removed all but one competitive district. And it took the two new representatives the state was apportioned in 2020 and drew two new congressional districts where Republicans are expected to win. This almost ensures the state’s congressional delegation will stay bright red for the next decade.
Although rosy for Democrats, the 2022 round of redistricting reflects the arms-race tendency of gerrymandering. “National partisan fairness is perfectly compatible”, Mr Stephanopolous says, “with extreme subnational partisan unfairness.” Bias in red states such as Texas and Florida, as well as blue states such as New Mexico and Oregon, still produces many wasted voters in congressional races, even if the number of seats parties are expected to win nationally is roughly fair. It is no puzzle who is hurt when politicians have the power to select their own voters.