N&O analysis confirms aggressive gerrymandering in NC political maps

The newest political maps from Republican lawmakers show strong evidence of partisan gerrymandering likely to help the GOP gain an edge in Congress and protect its majorities in the state legislature, according to a News & Observer analysis.

With conservative-dominated state and federal courts unlikely to rein in the worst abuses of political mapmaking, Republican gains from the maps — if approved — are likely to continue for years to come.

“It looks like, as we would have expected, that the deck is pretty stacked,” said Irving Joyner, an N.C. Central University law professor who’s been active in fights against racially discriminatory maps in this state for years.

The N&O analysis compared the performance of the proposed district maps to that of tens of thousands of maps created by a team of mathematicians at Duke University, using vote counts from previous elections.

The Duke team’s findings have been submitted to the U.S. Supreme Court and North Carolina state courts as evidence that previous political maps were extreme partisan gerrymanders. They were among the arguments that convinced state judges that previous Republican-drawn district maps violated the North Carolina constitution.

What the analysis found

The N&O analysis shows that among the 100,000 computer-generated maps, vanishingly few favor Republicans as much as the proposals now before state lawmakers, who have the power to pass them without the signature of Democratic Gov. Roy Cooper.

The Duke team confirmed The N&O’s findings last week, and in a blog post Friday said their own analysis found the maps even more gerrymandered than those Republicans drew in 2021.

That’s especially true of the two U.S. House map proposals, said Duke math professor Jonathan Mattingly, who leads the research team. No matter how the well Democrats perform, simulation after simulation shows almost no change in the makeup of the congressional delegation, reliably electing 10 or 11 Republicans compared to the current 7-7 party split.

“The maps the legislature have proposed essentially negate the need to have elections for the U.S. House of Representatives,” Mattingly said in an interview Friday.

Western Carolina political science professor Chris Cooper, a close observer of North Carolina politics who reviewed The N&O’s findings, said that although “there is no silver bullet in redistricting analysis,” he sees clear signals that the maps are particularly good for Republicans.

“They look fairly similar to maps that have been struck down in recent years by the courts,” Cooper said. “With the state supreme court we had two years ago, these maps may have crossed the line.”

But times have changed.

After Republican justices captured the majority of the state Supreme Court in the 2022 election, they quickly reversed course from previous decisions, declaring that partisan gerrymandering was a political issue, not a judicial one.

Republican leaders, who for the first time in several redistricting cycles created maps behind closed doors, contend the maps will withstand legal scrutiny. And they say the proposals give voters — not the courts — ample opportunity to determine the balance of political power.

“It doesn’t lock in any majority at all for the rest of the decade. Anybody who thinks that, they’re making a mistake,” Republican state Rep. Destin Hall, who chairs the House redistricting committee, told reporters Oct. 19. “We’ve still got to go out and campaign very hard. What will determine if we have majorities in this body will be the principles that we’ve passed throughout this decade and this session as well.”

2022 Congressional Map

2023 Proposed Congressional Map

Gerrymandered Districts

Partisan gerrymandering — whether it's by Democrats in Maryland or Republicans in North Carolina — can skew elections in favor of one party or another.

But detecting signs of partisan gerrymandering in political maps can be complicated, especially in a state like North Carolina.

It's not just about ugly shapes like these U.S. House districts created over years.

Even maps without jagged lines and bizarre shapes can give Republicans or Democrats a distinct advantage.

To find signs of partisan gerrymandering, we ideally need examples of maps considered "neutral," "normal." Or even "fair."

To learn how those "fair" maps behave, researchers at Duke University in 2021 used an algorithm to generate 100,000 possible maps of North Carolina districts.

Using election results, they added actual vote tallies — precinct by precinct — into each map.

By tallying up the results, they can show how many Republicans or Democrats each map would elect.

Votes cast in the 2020 U.S. Senate election, for example, would most commonly elect 55 Democrats in the state House.

That lets us compare possible maps with actual maps drawn by lawmakers.

The state House map lawmakers are now proposing would have elected 48 Democrats using the 2020 U.S. Senate election results.

Out of 100,000 maps the Duke team created, none ended up with that outcome.

But maybe that's a fluke.

We can check. By re-sorting the votes into the possible maps again from another election. Such as the 2020 race for president.

Yet again, Democrats perform worse with the proposed state House districts than with every single one of the 100,000 maps generated by Duke.

A different election — the 2020 race for governor — shows a similar result. Out of all 100,000 maps, zero perform like the one proposed by Republicans.

But the map proposal isn't always so out of whack with what we'd expect from this "normal" collection of maps.

With some older election results — the 2016 presidential race, for example — Democrats would still fare worse than about 84% of the possible maps.

Andy Taylor, professor of political science and director of the Free and Open Societies Project at N.C. State University, told The N&O after reviewing the paper’s analysis that he wasn’t surprised to see a Republican bias in the maps “given that legislators are human beings with political interests that computers don’t have.”

But Taylor, who served as an expert witness for Republican leaders during past redistricting cases, also noted that he’s skeptical of conclusions about voting patterns in the proposed districts far into the future — especially for a set of maps meant to last through 2030.

“That this is a road to the promised land for Republicans for the next 6 years and a death sentence for Democrats — that’s exaggerated,” Taylor said.

While he said it’s not hard for a political party to gerrymander in its favor, it’s much more difficult to ensure a biased map holds up to the chaos of the American electorate, which moves, grows and changes its mind based on candidates and issues of the day.

If mapmakers miscalculate, he said, they can end up creating a “dummymander” that can backfire when it fails to account for political changes.

“It’s hubris, I think, on the part of legislators who are sitting there rubbing their hands in glee that they know exactly what the election is going to be like in 2028,” Taylor said.

When swings don’t translate to seats

But Cooper takes the opposite view. The voting public in the U.S. has broadly hardened their opinions, he said, and communities have grown more solidly partisan as people have gravitated to places where others think — and vote — like themselves.

“American politics is increasingly — not decreasingly — predictable,” Cooper said.

He noted that opinion poll after opinion poll shows that voters — conservative or liberal — say they oppose partisan gerrymandering. But as long as voters agree with the party in power on more high-priority issues like abortion or tax policy, they’re often unwilling to penalize politicians for drawing unfair maps at the ballot box.

“It’s a great example where most people agree, but most people agree it’s not the most important issue,” Cooper said.

And that means partisan gerrymandering, whether by Democrats or Republicans, can feed a related problem in politics: it can insulate elected representatives from public opinion.

Cooper said The N&O’s analysis also shows that if voters’ positions do shift — a “blue wave,” for example — the Republican maps are particularly resistant to swings in political sentiment. Under the conditions of these maps, he said, it would take much larger swings in voting patterns to wrest, for example, the legislative majority from Republican control.

Partisan gerrymandering produces political districts intentionally skewed in favor of one political party.

But it can also generate maps resistant to changing political sentiment — think red or blue waves.

We can test this too, by simulating what would happen if a particular party — in this case, the Democrats — did better and better.

Here's how that analysis looks for North Carolina's newly proposed state Senate district map.

We start with the unadjusted results of the 2020 presidential election, when Republicans won 50.7% of the popular vote in North Carolina.

If we distribute the 2020 vote share using the newly proposed Senate district map, Republicans would have won 31 seats and a supermajority in the NC House.

But if we distribute that 2020 vote share using the most common election outcome across all 100,000 maps generated by researchers, Republicans would have only won 29 seats and a simple majority in the House.

Now, watch how the map drawn by lawmakers compares to the maps generated by researchers when the vote share shifts.

In each simulation, we can increase the Democratic vote share little by little…

... again and again.

The map collection drawn by algorithms — which used "neutral" criteria like equal population and county and city boundaries — for the most part follows changing voting patterns.

The map drawn by state lawmakers? A little less so.

It's a signal of partisan gerrymandering.

In N.C., lawmakers have incredible mapmaking power

In the end, strong evidence of partisan gerrymandering might not really matter, Taylor said.

Especially given the current state Supreme Court makeup, state lawmakers — and the party in charge — have incredible power to craft district lines.

“Of all the states in the union, the North Carolina constitution gives legislators more authority to do redistricting than any other state does — for better or worse,” Taylor said.

Even if tens of thousands of computer-generated map options appear more “fair,” he said, that’s unlikely to change the process. At least for now.

“It doesn’t matter what a Duke mathematician says,” Taylor said. “He or she has no constitutional authority at all.”

Whether courts step in or not, Cooper said there is value in deep dives into the maps to help the public see how their democracy works — or doesn’t

“It’s critically important for people to understand how their vote is likely to translate into representation,” Cooper said.

The N&O’s Kyle Ingram contributed to this report

How we reported this story

To analyze the districts proposed by Republican state lawmakers in October 2023 for signs of partisan gerrymandering, The News & Observer used a technique featured prominently in multiple legal challenges to maps across the country, including in North Carolina.

The analysis compares the proposals to a set of 100,000 computer-generated maps created by a team of mathematicians at Duke University led by Jonathan Mattingly. All of the maps in this "ensemble" follow neutral redistricting criteria established by the legislature during the 2021 redistricting process.

For each of these maps, the analysis uses precinct-level vote totals from past elections and re-sorts them into the different districts, simulating outcomes under different scenarios. The method doesn't predict how maps will perform in the future. Instead, it uses actual election data from past races unaffected by gerrymandering – president or governor, for example – where all voters cast ballots for the same set of candidates.

By counting how many Republicans or Democrats are elected using each map and in each election, the analysis can show how often a given proposal lines up with what we'd expect from the collection of "fair" maps.

To test how resistant the map proposals are to changing political sentiment, the analysis also simulated the outcomes if the Democratic vote percentage gradually increased.

For more on the analysis, see our full findings and methodology here.