I know the Illinois 13th looks artificial but, it groups a series of college and high population towns along a river valley that have different needs than the rural areas to the east and west.
I like Nate and respect his point of view. And that is why I am quite dejected when I read a piece from him that frames a political issue as more or less “Republicans are playing dirty and the Democrats are going to have to get dirty too as a result”. Too much NYT reading, Nate.
Democrats AND Republicans have been pushing the limits of gerrymandering for decades. I agree this is not good / healthy. I completely disagree that the republicans are the ones responsible for it happening. That is not supported by history at all.
Other than your dismissive blah blah blahs, you’re agreeing with me. The underlying reason that the republicans are the “conspicuous ones” at the moment is because they are more in power now than in recent history. They are more in power now because the Democrats decided to take positions that are extremely unpopular and, where and when in power, do a terrible job governing to help their own citizens. That has nothing to do with the Republicans playing dirty and causing the democrats to need to do the same.
I hope the Democrats find a way to push aside their far left activists so they can once again govern and bring some balance to our system. But so far they have not found a path to make it happen. In the meantime, the NYT, legacy media, and apparently sometimes Nate as well, will continue to live in their echo chamber where the Democrats are the good guys in their imaginary story.
You didn't look at the map in that analysis, let alone try to reason about it based on the material in Nate's article talking about the population per state.
It is pretty clear who is "pushing the limits of gerrymandering" right now. And Nate's article is about right now, not some hypothetical time that you falsely believe is still accurate.
And of course you are totally unbiased about claiming the D's "do a terrible job governing their own citizens". Funny thing is that "their own citizens" reelect the D's. Why do you hate democracy?
I looked at the link you shared—it’s analyzing the 2021 redistricting efforts. It seems you want to focus on “right now” rather than just history, though you also want to consider the historical context. You can point to the Princeton study, and I can point to analysis showing there are 35 Republican House districts in states with Democrat-controlled legislatures, but 55 Democrat House districts in states with Republican-controlled legislatures. So, who exactly is being the “bad actor” at this moment, or even in recent history?
In any case, it’s not straightforward to agree on a methodology for determining when a map is fair—or to create a spectrum or grading system to measure how unfair a map is. This is a major reason why state legislatures draw the maps in the first place: it’s inherently political. The court ruling Nate references was needed to clear up the confusion caused by prior decisions that required legislatures to consider race “a little,” but “not too much.” It became completely unworkable—no one knew whether a map was legal, leading to endless lawsuits and uncertainty.
When I criticized the Democrats earlier, it was based on two points I’d like your perspective on:
As a party, Democrats are historically unpopular.
Jurisdictions dominated by Democrats are generally losing population.
Fair enough—it’s still a leap to conclude they govern poorly. But those results could certainly lead a reasonable person to that conclusion.
I have no doubt—and fully support—the idea of an opposition party successfully challenging Republicans. It will happen, and likely sooner than many expect. But I believe it will come from non-Republicans who also reject the progressive extremists and “book-smart, real-world-moron” globalists who have, in my view, poisoned the Democratic Party over the last 10–15 years. They’re playing a game they cannot win because the world simply doesn’t work the way they imagine it does. They should spend more time studying reality and less time trying to force others to accept the flaws in their own worldview.
The Democrats controlled all three levers of power in 2021 and now control zero yet have the same basic leadership in Congress? The Party has had one really ope primary for president since 2008. So why should we be surprised when the minority leader won’t e dorse the Democratic nominee for Mayor? Remind me how old is Schumer?
Although the D's decision to keep the filibuster is part of the same "play nice" toolbox.
And the facts aren't on your side about 2016 or 2020.
Again, the D's play nice in more states and allow Independents to vote in the Democratic primaries. Bernie isn't even a footnote in history if the D's used the same process as the R's.
"But the bigger issue is that if both parties pursue a maximalist strategy, there’s no particular reason why Democrats should expect to be at a structural advantage."
Taking into account number of seats is a definite step forward from lazily just looking at number of states. I wonder if we could improve this metric further by taking into account the "potential" vs "realized" leverage of gerrymandering in a state relative to a baseline seat distribution which is based on the partisan lean of the state as measured by House popular vote, say.
For instance, based on the 2024 statewide popular vote across House seats, a purely proportional representation baseline in California would mean about 31.5 D to 20.5 R, which means the theoretically maximally ruthless gerrymander affords a potential 20ish seats, but the actual seat distribution is already more than halfway there, at 43 D to 9 R. So in fact there's only a potential 9 additional seats that could be "stolen" from gerrymandering.
Meanwhile in Texas, a 2024 popular vote baseline would yield 22.5 R seats to 15.5 D, providing a potential gerrymandering skew of 15 seats, whereas the actual current distribution is 25 R to 13 D -- much closer to the baseline than it is in California, and leaving a potential 13 seats to be "stolen" by gerrymandering. Texas has fewer seats in total, but there's more "untapped gerrymandering potential" there than in California.
Illinois "should be" a 9-8 split based on House popular vote, but is currently 14-3 -- like California, about half of the maximum potential gerrymandering impact is already baked in. Florida "should be" 16-12 in favor of Rs but is 20-8. So by this crude measure it "only" has 1/3 of its gerrymandering potential baked in.
Bottom line, even though Dem trifectas cover states accounting for about the same number of House seats as the ones where Republicans have trifectas, at least looking at the biggest of those they've already tapped more of that potential. Meaning Republicans may stand to gain more from this arms race going forward. (With the caveat that I haven't gone through every state, maybe the smaller ones add up to offset this) Not to mention that it's generally structurally harder to gerrymander in favor of a party whose voters tend to cluster in high density areas.
Trump is an authoritarian menace. Democrats need to control one House of Congress to provide some kind of check. Republicans are trying to cook the books. Cook the books back or be irrelevant. This is not hard.
No, no, NO! You are literally just feeding the fire and validating the Republican point that "both sides" do it. If Democrats are not going to take a principled stance on literally the topic of representation, then they don't deserve to have control either!
While gerrymandering may help the Democratic party, the rhetoric of protecting Democracy against rigging by Republicans repeats an error going back to 2016. Democrats are fighting for republican institutions--courts, rule of law, bureaucracy, traditional norms--against populist democracy. Gerrymandering is democratic assertion of control by elected legislatures over what courts, elites, experts, centerists and comfortable people like.
Calling every populist attempt to assert democratic power an existential threat to Democracy wore thin nearly a decade ago and repels thinking centerists, especially since in deep blue jurisdictions where Democrats win elections they seem to discard any respect for republican institutions and traditional norms. If people come to see the Democratic party strategy is stifling the populist efforts of the right until it can gain the power to unleash its own populist forces, they will not be inclined to support it.
Second, the first chart does not show 2024 having a more Gaussian distribution, it shows it having a smaller standard deviation. Rather than using a uniform distribution for the horizontal axis, if you want to evaluate how Gaussian a distribution is, you use a Q-Q plot with Gaussian axes. Rather than going from 1 to 435 equally spaced, you go from -3 to +3 standard deviations, spaced like a Gaussian curve.
When you do that, you see both 2016 and 2024 are quite similar in shape, mainly differing in that standard deviation was 12% higher in 2016. You also see the Democrats' disadvantage in the asymmetry. Democrats outperform Gaussian in almost all of one-third of the districts from +0.4 standard deviations up, underperforming for only a handful of the deepest blue ones. This does them little good since they would win all those districts anyway. Democrats also outperform in the one-sixth of districts from -1.0 standard deviations down, also doing little good since they lose all those districts even with the outperformance.
That leaves half the districts between -1.0 and +0.4 standard deviations--all the competitive places--in which Democrats underperform Gaussian in every single one, both in 2016 and 2024.
There's nothing magic about Gaussian, no reason to think the distribution should be Gaussian. Nevertheless, the party that outperforms in deep red and deep blue districts will necessarily underperform in competitive districts simply because the graphs are centered at zero. Whether the asymmetry is due to gerrymandering or something else, it could probably be addressed--at least in part--by Democratic gerrymandering--moving their "extra" voters out of districts where they do no good, into districts they might turn, or make safer.
Finally, I dispute that gerrymandering leads to whack-job candidates. If Republicans redraw boundaries so a 70% Republican district neighboring a 40% Republican district produces two 55% districts, they may gain a seat in Congress, but both Republicans will have to move toward the center to win, and both will be more vulnerable to demographic shifts and political trends. Gerrymandering can also concentrate voters, for example, if Democrats were in control they might take those two 55% Republican districts and restore them to 70% and 40% Republican to gain a seat. But that could allow a radical Republican to get into Congress--in other words, the whack jobs will be in the opposite party of the gerrymanderers.
I did a lot of statistical work in the 1980s trying to make juries fairer. One immediate observation is focusing on the percentage of Black versus White juries missed the point. All-white juries could delivered shockingly racist verdicts even in places thought to be quite liberal. A single Black juror completely changed deliberations. Additional Black jurors could even shift verdicts away from Black defendants. I came to believe that the fairest jury should include one member that matched the defendant on every possible dimension: race, sexual identity, age, income level, education, etc.
Similarly, labeling all voters as Republican or Democratic and counting them up misrepresents how electorates select and influence candidates. Even small groups of voters can exert significant influence on who gets nominated, who wins and how they act in office. That's why representatives will jettison ideology and national party desires to defend employers in their district or other matters that affect even a few hundred voters. To the extent gerrymanders mixes up groups so that every Congressional district contains a few hundred voters along every possible dimension, we could end up with a better Congress.
I love your observation about jury composition, but I also think we need to acknowledge some important differences between juries and electoral constituencies. Specifically, the impact of simple non-homogeneity is likely to be very different in a _deliberative_ body like a jury, where everyone sees, hears, and can directly interact with everyone else, than on an "aggregate" like an electorate.
If "minority" voters (not just in the race sense but in the numerical sense of representing a minority of the population along whatever categorical dimension) are distributed across districts, they may well wind up having almost no power, because in each individual election they are drowned out by the majority. Ensuring that every Congressional districts contains a perfectly representative cross-section of the population would actually produce the _worst_ gerrymander possible, by essentially ensuring that whatever viewpoint has a bare majority across the state is also a majority in every district, and therefore gets to be represented in every single seat.
Obviously that's a theoretical extreme, but the general point stands: there's value in having different districts have different compositions of viewpoints, because that allows minority viewpoints to be represented at the level of the deliberative body, where mere presence is likely to have more of an impact than it does at the level of the electorate.
I recognize there are differences between juries and Congressional districts, but I think the similarities are stronger than you suggest.
If you view elections in strictly partisan terms, then 49% Democratic voters are disenfranchised in a 51% Republican district. But there are many more dimensions of people, and different degrees of political attachment.
Suppose Blacks make up 10% of a Congressional district. I disagree that they will have "almost no power." For one thing, many people in the district will know Black people as neighbors, customers, servers, teachers, students and other capacities. For another, candidates will be careful not to drive Black voters away, since they could swing a close--or even not-so-close--election. The presence of a minority changes an entire culture.
This is even more true for less obvious minorities. Homosexuals may remain closeted if they are 1% of the population, but become more open if they are 5%. A few scattered veterans might not make much difference, but if there are enough for parades and a VFW or American Legion post they can be influential.
For this reason, I have always supported allowing incarcerated citizens to vote. If they could, candidates would visit prisons. If they did, I have little doubt prison conditions would improve dramatically. It's not that the incarcerated are a large voting bloc, but if they can't vote they will be ignored.
These are all good points that highlight what I think are very real "critical mass" effects in communities. That said, I don't see what it has to do with district boundaries. Presumably we're not talking about engineering districts by forcing people to move within a state; we're working with communities and social networks as they exist, and creating artificial political boundaries between districts for the purpose of electoral representation, which outside of those district-level elections, have little concrete "top-down" impact on who interacts with whom in real life, what kinds of parades are held, etc.
Definitely agree with you on disenfranchisement of the incarcerated, though.
Redistricting won't change people's neighbors or co-workers, but it does change Congressional candidates. With 10% Blacks in a district, both candidates will hold events in Black churches and say some nice things about Blacks on Juneteenth. Both will get their pictures taken at a soul food place (even if none of the Blacks in the district have connections with southern Black culture). Both will think about how their ads will play among Black voters. Both will have 10% Black campaign workers, including at least one senior staffer.
I think these experiences change candidates, or perhaps change who chooses to become a candidate.
It strikes me as odd to portray gerrymandering as a victory for populist democracy over republican institutional entrenchment, given that its effect tends to be making the composition of the legislative body _less responsive_ to the popular vote by reducing the number of competitive seats and protecting incumbents. I suppose you could argue it makes the parties _separately_ more populist insofar as it shifts the locus of electoral control toward primaries rather than the general election, but that's not the same as making the whole process more populist.
I don't claim gerrymandering makes democracy better. I only claim that allowing elected officials to draw boundaries is more democratic than letting unelected judges or commissions do it; or having rules that constrain legislative choices.
Democracy is not always good. The Constitution works as well as it does because it balances democracy with republican institutions, with checks and balances. Democracy holds the ultimate power--enough people can elect representatives to do what they want, including impeaching judges and installing new officials, and even amending the Constitution.
But the Constitution puts many hurdles in the path of pure democracy that slow it down and require long-lasting supermajorities to override courts and government departments.
Most people love democracy when they're in the majority, and yell about their rights or traditional norms or other alternatives when they're in the minority. That's understandable. What's hard to understand is people in the minority claiming they're fighting for democracy while opposing the results of the election they just lost.
I hope you're being facetious. I admit you won't find a lot of thoughtful centerists posting anonymous Internet rants or among media shock-jocks. But a small majority of people I know fall into the political center-right or center-left, and many of them are thoughtful.
Most are not very engaged politically and don't follow the general national news much, but they do have general impressions like the national Democratic party has been yelling "the sky is falling" for a decade and things don't seem so terrible; and that everything Republicans do, Democrats hate reflexively and in apocalyptic terms.
I don't mean the centrists I know have a positive view of the national Republican party, but they have different impressions about it than they do about the national Democratic party.
"I’m slightly sympathetic to the Roberts Court in the sense that I don’t think there’s any magic algorithm to ensure 'fair' districts ..."
I thought you were the data and stats guy Nate, what about using the efficiency gap to minimize wasted votes in districts? There are mathematical ways to ensure a fairer process than our current system of doing nothing.
While I agree with Nate's specific phrasing that there's no "magic algorithm" -- insofar as there are lots of metrics you could try to optimize and as with voting systems, none of them is going to strictly dominate all the others on every criterion you might care about -- I think there's a lot of value in democratically establishing a set of _criteria_ that the process is going to optimize rather than giving elected officials free rein to set arbitrary maps. This is basically the liberal part of liberal democracy: the public doesn't vote every time there's a decision to be made, even through their elected representatives; instead they vote on how the _system_ is set up, and have oversight over those responsible for faithfully carrying out the system as it's defined.
Nate, I recall you posting some versions of “fair ?” Districting years ago on 538. One was sort of an algorithm based on creating fairly normal geometric shaped districts with similar population (or densities?). Are you able to reference them or share the basic gist?
Anyway, always great, Nate. Enjoy Bergen. Hope you can get out to Geirangerfjord, or at least consider it. The best of Norway (my American opinion).
Just thinking, re. footnote 7: Why does either party even bother running candidates in the nearly 400 stone-cold-safe seats they have what amounts to zero chance of winning? (You could go back to at least the turn of the century and not find a single "upset" in seats rated "Safe" by 538 or Cook.) If only as a kind of "marketing" ploy to highlight the anti-democratic absurdity of the situation, I'd recommend to the Democrats at least that they seriously consider letting ALL safe Republican seats go uncontested -- and take whatever funding individual candidates or the party would have allocated to those useless campaigns and port it over to the few that are competitive.
Especially because the new max-gerrymandering will potentially put a few MORE seats in play. With the maps are so max-cynical, shouldn't campaign strategy and resource allocating be just as max-cynical to match? And let the voters get mad in the 90% of the districts where they have no choice - instead of going through the motions of fielding a doomed candidate to make it seem as if they do have a choice.
Poring over the sentence leads me to believe that the intended meaning is as follows. We define a marginal seat as every House seat in a US state after the first House seat in that state.
Sympathy for the Roberts Court is a foolish endeavour. They seem hell-bent on making the most obsequious concessions to the Hertigate Foundation's goals, then promptly are surprised with the Trump administration uses the new found freedoms beyond their intent, but in the most predicivale fashion. They are rather shocking stupid , or think we're incredibly gullible.
Picture a Democratic governor explaining: we had to suppress some of your votes, because Republicans suppressed votes in other states. That sounds un-American. Moderates won’t respect it—they’ll see Democrats embracing Trump’s tactics. Better to enact procedures so blue states can retaliate promptly when Republicans gerrymander. That would prove Democrats have the votes and the will to hit back, and understand that lawful vote suppression might be a necessary evil but really would rather avoid it.
Trump is a threat to democracy. This increases the incentive to bend the rules; it may lessen the odor of suppressing votes, but the stench is real and there will be a price to pay at election time. Roughly four House seats each cycle are decided by less than a point, so even a small loss in expected vote share could matter as much as gerrymandering Washington State and New York. Democracy will be safer if Democrats win the House.
I couldn't care one bit if this hurts or helps the Democratic Party. The fact that it makes our government less representative is bad, no matter how you cut it.
Also, pretending that "they go low, we go high" didn't work is absurd because you actually have to take the high road to even say you gave it an honest try. Democrats have chosen to go low but just go slightly less low (or do nothing) which is not the same as actually taking a principled stance.
"The initiative marks the end of a decade-plus of a “when they go low, we go high” attitude among Democratic leaders..."
If the past decade plus has been the Democrats taking the "high" road, what does it look like when they don't? Maps so blatantly skewed that even the NY Court of Appeals can't pretend they are fair? Launching investigations knowing full well that the evidence is completely fabricated? Character assassination? Using the power of government to pressure private businesses to take actions that negatively impact political opponents?
Oh wait, that's all stuff that they did when they were supposedly going "high."
Pretending that one side has cleaner hands than the other these days is either a sign of someone in denial, someone being intentionally obtuse, or someone who is hopelessly clueless. I know you're not clueless, and I'd hope you're not trying to deceive your readers. You need to stop looking at pyramids and get out of that river. The GOP has tossed out the people on their side who tried staying off the low road. The Democrats didn't have anyone fitting that description in the first place.
I know the Illinois 13th looks artificial but, it groups a series of college and high population towns along a river valley that have different needs than the rural areas to the east and west.
I like Nate and respect his point of view. And that is why I am quite dejected when I read a piece from him that frames a political issue as more or less “Republicans are playing dirty and the Democrats are going to have to get dirty too as a result”. Too much NYT reading, Nate.
Democrats AND Republicans have been pushing the limits of gerrymandering for decades. I agree this is not good / healthy. I completely disagree that the republicans are the ones responsible for it happening. That is not supported by history at all.
Blah blah blah. Both sides. Blah blah.
Here is some data to go with your "history":
https://gerrymander.princeton.edu/redistricting-report-card/
It may very well be that some in the Democratic party have been on the Gerrymander team in the past.
But right now the Republicans are the conspicuous ones.
Other than your dismissive blah blah blahs, you’re agreeing with me. The underlying reason that the republicans are the “conspicuous ones” at the moment is because they are more in power now than in recent history. They are more in power now because the Democrats decided to take positions that are extremely unpopular and, where and when in power, do a terrible job governing to help their own citizens. That has nothing to do with the Republicans playing dirty and causing the democrats to need to do the same.
I hope the Democrats find a way to push aside their far left activists so they can once again govern and bring some balance to our system. But so far they have not found a path to make it happen. In the meantime, the NYT, legacy media, and apparently sometimes Nate as well, will continue to live in their echo chamber where the Democrats are the good guys in their imaginary story.
You didn't look at the map in that analysis, let alone try to reason about it based on the material in Nate's article talking about the population per state.
It is pretty clear who is "pushing the limits of gerrymandering" right now. And Nate's article is about right now, not some hypothetical time that you falsely believe is still accurate.
And of course you are totally unbiased about claiming the D's "do a terrible job governing their own citizens". Funny thing is that "their own citizens" reelect the D's. Why do you hate democracy?
I looked at the link you shared—it’s analyzing the 2021 redistricting efforts. It seems you want to focus on “right now” rather than just history, though you also want to consider the historical context. You can point to the Princeton study, and I can point to analysis showing there are 35 Republican House districts in states with Democrat-controlled legislatures, but 55 Democrat House districts in states with Republican-controlled legislatures. So, who exactly is being the “bad actor” at this moment, or even in recent history?
In any case, it’s not straightforward to agree on a methodology for determining when a map is fair—or to create a spectrum or grading system to measure how unfair a map is. This is a major reason why state legislatures draw the maps in the first place: it’s inherently political. The court ruling Nate references was needed to clear up the confusion caused by prior decisions that required legislatures to consider race “a little,” but “not too much.” It became completely unworkable—no one knew whether a map was legal, leading to endless lawsuits and uncertainty.
When I criticized the Democrats earlier, it was based on two points I’d like your perspective on:
As a party, Democrats are historically unpopular.
Jurisdictions dominated by Democrats are generally losing population.
Fair enough—it’s still a leap to conclude they govern poorly. But those results could certainly lead a reasonable person to that conclusion.
I have no doubt—and fully support—the idea of an opposition party successfully challenging Republicans. It will happen, and likely sooner than many expect. But I believe it will come from non-Republicans who also reject the progressive extremists and “book-smart, real-world-moron” globalists who have, in my view, poisoned the Democratic Party over the last 10–15 years. They’re playing a game they cannot win because the world simply doesn’t work the way they imagine it does. They should spend more time studying reality and less time trying to force others to accept the flaws in their own worldview.
Shrug - Democrats group in major cities, which makes it tricky to completely exclude them without pizza pie slices into Republican rural areas.
https://www.nytimes.com/interactive/2025/08/21/upshot/up-massachusetts-redistricting.html
Popularity is a flawed question, since both the left and right are mad at the Dems right now.
As for population, right now, the states losing population are West Virginia, Louisiana, and Mississippi. What was your point again?
As for governing poorly, you will have to look far and wide to find any metric that shows D's are worse than R's by any objective measure.
Not understanding what's plainly obvious is a key halmark of the new republican party.
The Democrats controlled all three levers of power in 2021 and now control zero yet have the same basic leadership in Congress? The Party has had one really ope primary for president since 2008. So why should we be surprised when the minority leader won’t e dorse the Democratic nominee for Mayor? Remind me how old is Schumer?
N key giving you troubles, Gary?
Redistricting is the topic.
Although the D's decision to keep the filibuster is part of the same "play nice" toolbox.
And the facts aren't on your side about 2016 or 2020.
Again, the D's play nice in more states and allow Independents to vote in the Democratic primaries. Bernie isn't even a footnote in history if the D's used the same process as the R's.
"But the bigger issue is that if both parties pursue a maximalist strategy, there’s no particular reason why Democrats should expect to be at a structural advantage."
Is this supposed to be "disadvantage"?
That was my question!
Taking into account number of seats is a definite step forward from lazily just looking at number of states. I wonder if we could improve this metric further by taking into account the "potential" vs "realized" leverage of gerrymandering in a state relative to a baseline seat distribution which is based on the partisan lean of the state as measured by House popular vote, say.
For instance, based on the 2024 statewide popular vote across House seats, a purely proportional representation baseline in California would mean about 31.5 D to 20.5 R, which means the theoretically maximally ruthless gerrymander affords a potential 20ish seats, but the actual seat distribution is already more than halfway there, at 43 D to 9 R. So in fact there's only a potential 9 additional seats that could be "stolen" from gerrymandering.
Meanwhile in Texas, a 2024 popular vote baseline would yield 22.5 R seats to 15.5 D, providing a potential gerrymandering skew of 15 seats, whereas the actual current distribution is 25 R to 13 D -- much closer to the baseline than it is in California, and leaving a potential 13 seats to be "stolen" by gerrymandering. Texas has fewer seats in total, but there's more "untapped gerrymandering potential" there than in California.
Illinois "should be" a 9-8 split based on House popular vote, but is currently 14-3 -- like California, about half of the maximum potential gerrymandering impact is already baked in. Florida "should be" 16-12 in favor of Rs but is 20-8. So by this crude measure it "only" has 1/3 of its gerrymandering potential baked in.
Bottom line, even though Dem trifectas cover states accounting for about the same number of House seats as the ones where Republicans have trifectas, at least looking at the biggest of those they've already tapped more of that potential. Meaning Republicans may stand to gain more from this arms race going forward. (With the caveat that I haven't gone through every state, maybe the smaller ones add up to offset this) Not to mention that it's generally structurally harder to gerrymander in favor of a party whose voters tend to cluster in high density areas.
Trump is an authoritarian menace. Democrats need to control one House of Congress to provide some kind of check. Republicans are trying to cook the books. Cook the books back or be irrelevant. This is not hard.
No, no, NO! You are literally just feeding the fire and validating the Republican point that "both sides" do it. If Democrats are not going to take a principled stance on literally the topic of representation, then they don't deserve to have control either!
My dude. This thing is full of typos, from the title down. Slow down and proofread a little before you post something.
Aside from that, interesting stuff.
Three comments.
While gerrymandering may help the Democratic party, the rhetoric of protecting Democracy against rigging by Republicans repeats an error going back to 2016. Democrats are fighting for republican institutions--courts, rule of law, bureaucracy, traditional norms--against populist democracy. Gerrymandering is democratic assertion of control by elected legislatures over what courts, elites, experts, centerists and comfortable people like.
Calling every populist attempt to assert democratic power an existential threat to Democracy wore thin nearly a decade ago and repels thinking centerists, especially since in deep blue jurisdictions where Democrats win elections they seem to discard any respect for republican institutions and traditional norms. If people come to see the Democratic party strategy is stifling the populist efforts of the right until it can gain the power to unleash its own populist forces, they will not be inclined to support it.
Second, the first chart does not show 2024 having a more Gaussian distribution, it shows it having a smaller standard deviation. Rather than using a uniform distribution for the horizontal axis, if you want to evaluate how Gaussian a distribution is, you use a Q-Q plot with Gaussian axes. Rather than going from 1 to 435 equally spaced, you go from -3 to +3 standard deviations, spaced like a Gaussian curve.
When you do that, you see both 2016 and 2024 are quite similar in shape, mainly differing in that standard deviation was 12% higher in 2016. You also see the Democrats' disadvantage in the asymmetry. Democrats outperform Gaussian in almost all of one-third of the districts from +0.4 standard deviations up, underperforming for only a handful of the deepest blue ones. This does them little good since they would win all those districts anyway. Democrats also outperform in the one-sixth of districts from -1.0 standard deviations down, also doing little good since they lose all those districts even with the outperformance.
That leaves half the districts between -1.0 and +0.4 standard deviations--all the competitive places--in which Democrats underperform Gaussian in every single one, both in 2016 and 2024.
There's nothing magic about Gaussian, no reason to think the distribution should be Gaussian. Nevertheless, the party that outperforms in deep red and deep blue districts will necessarily underperform in competitive districts simply because the graphs are centered at zero. Whether the asymmetry is due to gerrymandering or something else, it could probably be addressed--at least in part--by Democratic gerrymandering--moving their "extra" voters out of districts where they do no good, into districts they might turn, or make safer.
Finally, I dispute that gerrymandering leads to whack-job candidates. If Republicans redraw boundaries so a 70% Republican district neighboring a 40% Republican district produces two 55% districts, they may gain a seat in Congress, but both Republicans will have to move toward the center to win, and both will be more vulnerable to demographic shifts and political trends. Gerrymandering can also concentrate voters, for example, if Democrats were in control they might take those two 55% Republican districts and restore them to 70% and 40% Republican to gain a seat. But that could allow a radical Republican to get into Congress--in other words, the whack jobs will be in the opposite party of the gerrymanderers.
I did a lot of statistical work in the 1980s trying to make juries fairer. One immediate observation is focusing on the percentage of Black versus White juries missed the point. All-white juries could delivered shockingly racist verdicts even in places thought to be quite liberal. A single Black juror completely changed deliberations. Additional Black jurors could even shift verdicts away from Black defendants. I came to believe that the fairest jury should include one member that matched the defendant on every possible dimension: race, sexual identity, age, income level, education, etc.
Similarly, labeling all voters as Republican or Democratic and counting them up misrepresents how electorates select and influence candidates. Even small groups of voters can exert significant influence on who gets nominated, who wins and how they act in office. That's why representatives will jettison ideology and national party desires to defend employers in their district or other matters that affect even a few hundred voters. To the extent gerrymanders mixes up groups so that every Congressional district contains a few hundred voters along every possible dimension, we could end up with a better Congress.
I love your observation about jury composition, but I also think we need to acknowledge some important differences between juries and electoral constituencies. Specifically, the impact of simple non-homogeneity is likely to be very different in a _deliberative_ body like a jury, where everyone sees, hears, and can directly interact with everyone else, than on an "aggregate" like an electorate.
If "minority" voters (not just in the race sense but in the numerical sense of representing a minority of the population along whatever categorical dimension) are distributed across districts, they may well wind up having almost no power, because in each individual election they are drowned out by the majority. Ensuring that every Congressional districts contains a perfectly representative cross-section of the population would actually produce the _worst_ gerrymander possible, by essentially ensuring that whatever viewpoint has a bare majority across the state is also a majority in every district, and therefore gets to be represented in every single seat.
Obviously that's a theoretical extreme, but the general point stands: there's value in having different districts have different compositions of viewpoints, because that allows minority viewpoints to be represented at the level of the deliberative body, where mere presence is likely to have more of an impact than it does at the level of the electorate.
Thank you for the kind words.
I recognize there are differences between juries and Congressional districts, but I think the similarities are stronger than you suggest.
If you view elections in strictly partisan terms, then 49% Democratic voters are disenfranchised in a 51% Republican district. But there are many more dimensions of people, and different degrees of political attachment.
Suppose Blacks make up 10% of a Congressional district. I disagree that they will have "almost no power." For one thing, many people in the district will know Black people as neighbors, customers, servers, teachers, students and other capacities. For another, candidates will be careful not to drive Black voters away, since they could swing a close--or even not-so-close--election. The presence of a minority changes an entire culture.
This is even more true for less obvious minorities. Homosexuals may remain closeted if they are 1% of the population, but become more open if they are 5%. A few scattered veterans might not make much difference, but if there are enough for parades and a VFW or American Legion post they can be influential.
For this reason, I have always supported allowing incarcerated citizens to vote. If they could, candidates would visit prisons. If they did, I have little doubt prison conditions would improve dramatically. It's not that the incarcerated are a large voting bloc, but if they can't vote they will be ignored.
These are all good points that highlight what I think are very real "critical mass" effects in communities. That said, I don't see what it has to do with district boundaries. Presumably we're not talking about engineering districts by forcing people to move within a state; we're working with communities and social networks as they exist, and creating artificial political boundaries between districts for the purpose of electoral representation, which outside of those district-level elections, have little concrete "top-down" impact on who interacts with whom in real life, what kinds of parades are held, etc.
Definitely agree with you on disenfranchisement of the incarcerated, though.
Yes, I got a little off-path in my first reply.
Redistricting won't change people's neighbors or co-workers, but it does change Congressional candidates. With 10% Blacks in a district, both candidates will hold events in Black churches and say some nice things about Blacks on Juneteenth. Both will get their pictures taken at a soul food place (even if none of the Blacks in the district have connections with southern Black culture). Both will think about how their ads will play among Black voters. Both will have 10% Black campaign workers, including at least one senior staffer.
I think these experiences change candidates, or perhaps change who chooses to become a candidate.
It strikes me as odd to portray gerrymandering as a victory for populist democracy over republican institutional entrenchment, given that its effect tends to be making the composition of the legislative body _less responsive_ to the popular vote by reducing the number of competitive seats and protecting incumbents. I suppose you could argue it makes the parties _separately_ more populist insofar as it shifts the locus of electoral control toward primaries rather than the general election, but that's not the same as making the whole process more populist.
I don't claim gerrymandering makes democracy better. I only claim that allowing elected officials to draw boundaries is more democratic than letting unelected judges or commissions do it; or having rules that constrain legislative choices.
Democracy is not always good. The Constitution works as well as it does because it balances democracy with republican institutions, with checks and balances. Democracy holds the ultimate power--enough people can elect representatives to do what they want, including impeaching judges and installing new officials, and even amending the Constitution.
But the Constitution puts many hurdles in the path of pure democracy that slow it down and require long-lasting supermajorities to override courts and government departments.
Most people love democracy when they're in the majority, and yell about their rights or traditional norms or other alternatives when they're in the minority. That's understandable. What's hard to understand is people in the minority claiming they're fighting for democracy while opposing the results of the election they just lost.
A thinking centrist? Honestly, I have no idea what this is supposed to mean in today's political climate.
I hope you're being facetious. I admit you won't find a lot of thoughtful centerists posting anonymous Internet rants or among media shock-jocks. But a small majority of people I know fall into the political center-right or center-left, and many of them are thoughtful.
Most are not very engaged politically and don't follow the general national news much, but they do have general impressions like the national Democratic party has been yelling "the sky is falling" for a decade and things don't seem so terrible; and that everything Republicans do, Democrats hate reflexively and in apocalyptic terms.
I don't mean the centrists I know have a positive view of the national Republican party, but they have different impressions about it than they do about the national Democratic party.
Fair, I understand how disliking the parties and being generally unplugged from national news could lead one to identify as a centrist.
Yes, those of us who like neither Team Red nor Team Blue never have a single thought…
I don’t like the teams either, but I’m not sure what it means to be in the center of them.
If the Democrats want to play hardball, it would be mean they have some balls. They haven’t any balls for decades.
"I’m slightly sympathetic to the Roberts Court in the sense that I don’t think there’s any magic algorithm to ensure 'fair' districts ..."
I thought you were the data and stats guy Nate, what about using the efficiency gap to minimize wasted votes in districts? There are mathematical ways to ensure a fairer process than our current system of doing nothing.
While I agree with Nate's specific phrasing that there's no "magic algorithm" -- insofar as there are lots of metrics you could try to optimize and as with voting systems, none of them is going to strictly dominate all the others on every criterion you might care about -- I think there's a lot of value in democratically establishing a set of _criteria_ that the process is going to optimize rather than giving elected officials free rein to set arbitrary maps. This is basically the liberal part of liberal democracy: the public doesn't vote every time there's a decision to be made, even through their elected representatives; instead they vote on how the _system_ is set up, and have oversight over those responsible for faithfully carrying out the system as it's defined.
Nate, I recall you posting some versions of “fair ?” Districting years ago on 538. One was sort of an algorithm based on creating fairly normal geometric shaped districts with similar population (or densities?). Are you able to reference them or share the basic gist?
Anyway, always great, Nate. Enjoy Bergen. Hope you can get out to Geirangerfjord, or at least consider it. The best of Norway (my American opinion).
Just thinking, re. footnote 7: Why does either party even bother running candidates in the nearly 400 stone-cold-safe seats they have what amounts to zero chance of winning? (You could go back to at least the turn of the century and not find a single "upset" in seats rated "Safe" by 538 or Cook.) If only as a kind of "marketing" ploy to highlight the anti-democratic absurdity of the situation, I'd recommend to the Democrats at least that they seriously consider letting ALL safe Republican seats go uncontested -- and take whatever funding individual candidates or the party would have allocated to those useless campaigns and port it over to the few that are competitive.
Especially because the new max-gerrymandering will potentially put a few MORE seats in play. With the maps are so max-cynical, shouldn't campaign strategy and resource allocating be just as max-cynical to match? And let the voters get mad in the 90% of the districts where they have no choice - instead of going through the motions of fielding a doomed candidate to make it seem as if they do have a choice.
But then the incumbents also don’t need to spend and they can potentially reallocate their funding as well. So it’s not clear this strategy is better.
> a “trifecta” — which I define simply as one less than seat its overall number apportioned in the 2020 Census
Say again?
Poring over the sentence leads me to believe that the intended meaning is as follows. We define a marginal seat as every House seat in a US state after the first House seat in that state.
Also, the graph makes that intended meaning clear.
Sympathy for the Roberts Court is a foolish endeavour. They seem hell-bent on making the most obsequious concessions to the Hertigate Foundation's goals, then promptly are surprised with the Trump administration uses the new found freedoms beyond their intent, but in the most predicivale fashion. They are rather shocking stupid , or think we're incredibly gullible.
Picture a Democratic governor explaining: we had to suppress some of your votes, because Republicans suppressed votes in other states. That sounds un-American. Moderates won’t respect it—they’ll see Democrats embracing Trump’s tactics. Better to enact procedures so blue states can retaliate promptly when Republicans gerrymander. That would prove Democrats have the votes and the will to hit back, and understand that lawful vote suppression might be a necessary evil but really would rather avoid it.
Trump is a threat to democracy. This increases the incentive to bend the rules; it may lessen the odor of suppressing votes, but the stench is real and there will be a price to pay at election time. Roughly four House seats each cycle are decided by less than a point, so even a small loss in expected vote share could matter as much as gerrymandering Washington State and New York. Democracy will be safer if Democrats win the House.
I couldn't care one bit if this hurts or helps the Democratic Party. The fact that it makes our government less representative is bad, no matter how you cut it.
Also, pretending that "they go low, we go high" didn't work is absurd because you actually have to take the high road to even say you gave it an honest try. Democrats have chosen to go low but just go slightly less low (or do nothing) which is not the same as actually taking a principled stance.
"The initiative marks the end of a decade-plus of a “when they go low, we go high” attitude among Democratic leaders..."
If the past decade plus has been the Democrats taking the "high" road, what does it look like when they don't? Maps so blatantly skewed that even the NY Court of Appeals can't pretend they are fair? Launching investigations knowing full well that the evidence is completely fabricated? Character assassination? Using the power of government to pressure private businesses to take actions that negatively impact political opponents?
Oh wait, that's all stuff that they did when they were supposedly going "high."
Pretending that one side has cleaner hands than the other these days is either a sign of someone in denial, someone being intentionally obtuse, or someone who is hopelessly clueless. I know you're not clueless, and I'd hope you're not trying to deceive your readers. You need to stop looking at pyramids and get out of that river. The GOP has tossed out the people on their side who tried staying off the low road. The Democrats didn't have anyone fitting that description in the first place.