MathIntee Political Post

Against antisemitism in America

Tomas McIntee — Sat, 06 Jun 2026 15:21:22 GMT

When I started writing this article the latest political antisemitism scandal to make national news was that Rama Duwaji, the wife of New York City’s new mayor, multi-million follower social media influencer, and the frequent subject of glowing puff pieces in the New York Times,1 liked and shared Hamas propaganda posts while the October 7th attacks were underway.2 Shortly after, violent antisemitic attacks displaced her in the news cycle.3

In her professional capacity as an artist, Rama Duwaji also recently drew an illustration (above, fair use without permission) for a notorious antisemitic activist’s “essay” (likely more appropriately called “short story”) titled “A Trail of Soap.” This became part of the same scandal.

This was not an isolated event for Zohran Mamdani, who supported a pro-Hamas fundraising group,4 has close ties to many overtly antisemitic political activists,5 and whose father defended suicide bombers in a book.6 His primary political passion is the Palestinian movement, which is full of antisemites.7 He hosted to a noted supporter of terrorism in the mayoral mansion a day after terrorists attacked anti-Islam protestors just outside.8

Nor is Zohran Mamdani the only Democrat with antisemitic associations; the week before Rama Duwaji’s social media history became news, Graham Platner, a frontrunner Senate candidate previously noteworthy for his Totenkopf tattoo,9 spurred controversy by declaring himself a long-time fan of an antisemitic conspiracy theorist.10

Nor is this a problem limited to Democrats. The entire Horseshoe Caucus has antisemitic tendencies shared across a bipartisan fringe. The Republican side of the political horseshoe includes politicians like Marjorie Taylor “Jewish Space Lasers” Greene and social media influencers like Nick Fuentes.11 Antisemitism has grown with polarization and radicazlization.

American philosemitism

In 1790, the American population was about 1% Catholic and less than 0.1% Jewish. As far as American sectarianism was concerned, Jews were not appreciably stranger in their religious practices than Catholics, Mormons, American Indians, or Orthodox Christians, and were barely visible. Immigration from continental Europe brought the Jewish population from ~0.2% by around 1848 to ~0.5% in 1877.

A positive portrayal of America welcoming Jewish immigrants from the magazine Puck in 1881. (Uncle Sam is a traditional male personification of the United States.)

From 1877 to 1900, this tradition of religious tolerance attracted enough Jews for America’s Jewish population to grow from negligible to the second-largest in the world. After World War I, the United States had the largest Jewish population in the world.12 After World War II, America had a larger Jewish population than all of Europe combined,13 though this was still only 3% of the total American population.

Another positive portrayal of America welcoming Jewish immigrants from 1881, this one in Frank Leslie’s Illustrated Newspaper. (Columbia is a traditional female personification of the United States.)

At present, about 2% of America’s population is Jewish, which makes America the second-most Jewish country of the 21st century on a per-capita basis.14 In absolute terms, less than 20% of the global Jewish population lives outside of Israel and the United States, with comparable shares in each country. This is a consequence of the United States being effectively one of the most philosemitic countries in the world for most of its existence.

A first wave of American antisemitism

Distinctly American antisemitism didn’t really get off the ground until the interwar period between World War I and World War II.15 In earlier periods, the Jewish population was very small, and nativist groups tended to focus more on Catholics than Jews. The Jewish population and nativist backlash against immigration both reached critical heights after World War I. This backlash had prominent supporters. Henry Ford blamed the Jews for World War I; he began publishing The International Jew in 1920.

Henry Ford and his newspaper. Public domain images.

In 1922, Harvard, followed by the rest of the Ivy League, began to discriminate against Jewish applicants. In 1926, Father Coughlin started a radio show to eliminate anti-Catholic prejudice; in 1930, it became a national show with overtly antisemitic messages alongside his pro-Catholic material. In 1936, he added print to his portfolio with a magazine called Social Justice.

Father Coughlin orating. Public domain image.

Father Coughlin and Henry Ford were sympathetic to Hitler early on; the broader American public was not, including broadcasters and Coughlin’s superiors in the Catholic hierarchy. Coughlin’s radio program was canceled in 1939, and he was forced to stop publishing Social Justice in 1942.

A second wave of American antisemitism

The next major venture in popular American antisemitism would come from the African-American community, where political activists were divided into integrationists (such as Martin Luther King Jr.) and separatists (such as Malcolm X). The integrationists mostly had friendly attitudes towards Jews.16

Both groups looked for examples outside of the United States for inspiration. Integrationists built a playbook based on the successful peaceful protest movement led by Gandhi. The separatists tended to align with international communists and Islamists. Black nationalists exchanged ideas and tactics with antisemitic Palestinians, antisemitic Algerians, et cetera. Many black nationalist leaders and organizations have been antisemitic, most notably the Nation of Islam.17

Both white nationalists (who usually regard Jews as non-white) and black nationalists (who usually regard Jews as white) are usually antisemitic, in addition to having a common cause of racial segregation.

White nationalists continued to be antisemitic, though they had less interaction with mainstream politicians after FDR sidelined Henry Ford and Father Coughlin.18 For example, a rising young Barack Obama met with Louis Farrakhan in 2005,19 and supporters of Farrakhan were involved with the anti-Trump Women’s March, which was considered politically mainstream.

A third wave of American antisemitism

Many left-horseshoe antisemites are first or second generation immigrants and disproportionately Muslim;20 many right-horseshoe antisemites are second or third generation immigrants and disproportionately Hispanic.21 Third-wave American antisemitism is as American as shawarma stands and taco trucks.22

This is not to say that first- and second-wave antisemitism were independent of international antisemitism,23 but they grew under the leadership of American (or Canadian-American24) figures as part of larger and fundamentally unrelated political projects.25 Third-wave antisemitism is more directly imported, with more foreign leadership.26 Its primary associated political projects are opposition to Israel and the decentralization of industries seen as controlled by Jews,27 with occasional side orders of the white and black racial nationalisms associated with the remnants of the two waves of American antisemitism.28

What lies ahead

The unfortunate reality is that all evidence suggests the third wave of American antisemitism is growing. Left-horseshoe antisemitism is closely linked to rising political violence. YouGov polls rank Zohran Mamdani as one of the top 10 most popular Democratic politicians in the country. The most prominent Democratic news media organizations in the country offer consistently positive coverage of third-wave antisemites on the basis of either political progressivism or immigration status.29

At the same time, broad left-wing efforts to censor and deplatform all right-wing figures have backfired by fueling the growth of alternative media platforms where right-wing antisemitic content is common.30 While the risk is less immediate on the Republican side of the aisle, post-Trump Republicans may fully embrace right-horseshoe antisemitism in the future.31

Broadly speaking, I am not optimistic, but the first step to solving the problem is identifying its existence.

C.f. 1 2 3 4 5 6 7. Not just the NYT, either, she was covered glowingly by many prominent magazines, such as People, Harper’s Bazaar, Elle, etc as “New York City’s new First Lady,” in a manner not typical for NYC mayor spouses but in keeping with the generally glowing coverage of Mamdani by Democratic-aligned national news media.

Which even the NYT has chosen to cover here, and less sympathetically by the Times of Israel and the New York Post. Deeper subsequent dives are also out there.

There was a 152% spike in the rate of hate crimes in New York City for January, mainly antisemitic hate crimes, which made the news because the NYPD has decided to revise its hate crime reporting methods in what I can only see as an attempt to drive the numbers downwards and prevent comparisons. Then on March 12th, there was a truck ramming attack against a synagogue in Michigan. Rama Duwaji’s social media posts have been further investigated since then by independent journalists, but have received little mainstream new coverage.

See here for discussion of “The Holy Land Five.”

It’s been a perennial news story. One installment here.

Mamdani’s father’s “research” (really, the academic equivalent of opinion-mongering) is largely in line with the fashionable and empirically incorrect postmodern school of thought that glorifies political violence and confuses activism with scholarship, like many related schools of activism.

While critique of the Israeli government is not intrinsically antisemitic, antisemites are generally reflexively hostile to Israel. Since most Americans aren’t passionate about Middle Eastern politics and most of the ones who do side with Israel, the anti-Israel movement is small enough to be dominated by the equally-fringe antisemites.

As discussed here. Whatever you might believe about the legality of the Trump administration’s efforts to deport Khalil, it is a pointed matter of fact that he is a Hamas-affiliated activist who has repeatedly endorsed and celebrated terrorism.

See here, but also I recommend reading this Twitter thread, which puts the tattoo in the context of USMC policy changes and Platner’s service history.

Some coverage here.

Discussion of Fuentes here.

A notable part of this was that the Russian Empire became the Soviet Union and lost a significant portion of its western territories.

Largely due to the Holocaust, with the difference rapidly becoming larger due to postwar migration to the United States and Israel.

While many other countries once had a 1% or higher Jewish share of their population, only the loosest possible definitions of “Jewish-connected” population put that share above 1% for Canada, France, and Hungary, and that same loose definition brings the US share up to 3%.

It is possible to find individual Americans making antisemitic statements before World War I, but there really isn’t that much.

For example, MLK had close Jewish associates, and was a staunch supporter of Israel.

Nation of Islam was founded in 1930, but did not become prominent until the 1950s. The Black Hebrew Israelites, an African-American group noted for its members’ participation in antisemitic violence, can trace their roots back to 1886, but they have remained a fringe group throughout without the kind of public promenince achieved by Henry Ford.

David Duke, perhaps the most prominent white antisemite of the latter half of the 20th century, lost his closest election by 10% in 1990. The next year, he lost a bid for governor of Louisiana by a more decisive margin of 22% against a known criminal in the famous “Klan versus Crook” election. Arguably, David Duke’s white nationalism was held back by his overt antisemitism and interests in anti-Israeli foreign policy; some of his political advisors tried to get him to focus more narrowly on being an anti-black bigot, as Louisiana’s local racists were mostly uninterested in antisemitism and foreign affairs, but they largely failed.

For a pointed contrast in the relative importance of black nationalists and white nationalists in the latter part of the 20th century, see Farrakhan’s “Million Man March” in contrast to the tens of thousands of votes David Duke received when running for president, but also social media, where both relics of 20th century racial nationalism continued to participate in public discourse into the 21st century: Farrakhan currently has 360,000 followers on Twitter, while David Duke had 32,000 followers before his suspension in 2017. This is especially striking given that African-Americans only make up around one-eighth of the US population.

See here.

See previous commentary about Zohran Mamdani’s many antisemitic associates, but also the encampent movement on US college campuses during the recent war in Gaza, which were to a large degree populated by international students and frequently involved antisemitic chants in Arabic.

Note the surnames of the Young Republicans in this article. Surveys indicate that Hispanic and Latino respondents are generally more antisemitic than the general American population, and especially the Republicans.

If you feel like arguing against this metaphor, I’ll point out a few things first. Some taco trucks are owned and operated by Mayflower-descended Anglo-Americans who love tacos; some are operated by unassimilated illegal immigrants. Some taco trucks offer a very autentico experience; others are fully Americanized versions of the cuisine. Similarly, third-wave antisemitism identifiably draws its ideas from the stock of Old World antisemitism, with a mixture of imported practitioners and indigenous innovators & imitators.

The Protocols of the Elders of Zion was circulated by Henry Ford’s magazine, but originated in Russia. Also as noted earlier, black nationalists exchanged ideas and inspiration with Palestinian and Algerian Islamists and other Islamist revolutionaries.

Father Coughlin was born in Ontario to an Irish-American father and an Irish-Canadian mother.

Henry Ford was also interested heavily in world peace, eugenics, and a leading American isolationist. Father Coughlin’s primary political project was initially fighting for the integration and acceptance of Catholics into mainstream American society. The Nation of Islam is primarily a black nationalist organization.

Zohran Mamdani is a naturalized citizen. Rama Duwaji, while born in the United States, is a second-generation immigrant who described herself as “from Damascus” on social media. Mahmoud Khalil, one of the main leaders of the Columbia encampment, faces deportation. Ilhan Omar is a naturalized citizen; Rashida Tlaib is a second-generation immigrant who puts her Palestinian identity front and center.

Namely banking and news media.

Black nationalism in particular experienced a large resurgence as related racial radicals became mainstream, though they may have peaked (“peak woke”) with the surge in interest in the Palestinian cause on the left. White nationalists are slowly tiptoeing back into view, but are still struggling with debanking and deplatforming along with overtly antisemitic pro-Trump African-Americans.

Mahmoud Khalil falls neatly into the overlap here.

For example, Nick Fuentes was banned on most platforms 2019-2020, joining Rumble in 2021. However, since many more normal right-wing figures were also being deplatformed with algorithmic suppression and joining Rumble at the same time, Rumble was rapidly growing in 2021 and this ended up increasing Fuentes’s visibility and following.

In spite of a consistent drumbeat of claims that Trump is antisemitic based on tenuous associations, Trump’s inner circle is disproportionately Jewish, his favorite daughter converted to Judaism, he kicked Marjorie “Jewish Space Lasers” Tayor Greene out of the party, he banned Kanye West from attending his 2025 inauguration, he backed Israel to the hilt in wars against Gaza and Iran, and while he has expressed some stereotypical views about Jews, he seems to generally regard those stereotypes as reflecting positively on Jews.

How to end the gerrymandering wars

Tomas McIntee — Wed, 13 May 2026 21:52:46 GMT

The Supreme Court threw out Louisiana’s racially gerrymandered map, clearing the way for Louisiana Republicans to use a stronger partisan gerrymander. At no point was there any possibility of a fair map for Louisiana; the case was about deciding between different kinds of gerrymandering. Alito celebrated partisan gerrymandering as a legitimate goal, building on Roberts’s opinion in Rucho, while limiting racial gerrymandering.

Louisiana’s racially gerrymandered districts, which will be replaced soon with a severe partisan gerrymander. Map (and others of similar style) by user TwoTwoFortySix on Wikimedia and under CC.

Partisan gerrymandering is toxic for democracy. The current legal path to ending partisan gerrymandering, also noted by Roberts in Rucho, lies through an act of Congress. In that light, I would like to outline a solution:

Force the drawing of districts to take place using objective standards to make it difficult to draw strong gerrymanders.
Implement a form of mixed-member proportional representation.
Expand the House to the cube root of the population.1

Each of these steps, by itself, helps reduce the impact of gerrymandering. Taken together, they would eliminate the problem with a net minimum of complications, offering a smooth transition to a better democratic process.

A Model Act to End Gerrymandering

Districts for the House of Representatives shall be compact, convex, contiguous, of equal population to within precision of measurement, and minimize divisions of county-level areas (“counties”) and other areas as defined and utilized within the previous decennial census. These requirements shall take effect upon enactment and apply to all redistricting conducted after this date. In particular:
1. Counties with population meeting to within precision of measurement or exceeding the state population divided by the number of districts (“district population quota”) shall, respectively, be a district, or completely contain a number of districts equal to the county population divided by the district population quota, rounded down.2
2. For counties with population exceeding the district population quota, the portion not assigned to a district completely contained within that county shall be assigned completely to a single district whenever possible.3
3. Districts shall be as contiguous as possible, which is to say that they shall be formed from counties or partial counties that are strongly adjacent when possible, and weakly adjacent when this is required by the geography of the state and the necessity of forming districts of equal population.
  1. A pair of counties or partial counties are strongly adjacent if they are connected by the whole of the land boundary between the two counties, which spans a distance from point to point exceeding one-tenth of the smaller perimeter of the convex hull of the two counties.4
  2. A partial county is also strongly adjacent to a county if the partial county is part of a county that is strongly adjacent and no part of the partial county is greater in distance from the boundary than the distance across the partial county’s boundary from point to point.5
  3. Otherwise, they are weakly adjacent if connected by a smaller land boundary or if they include the pair of closest points on two separate land masses that share no land boundary within the state.6
4. Except where doing so would conflict with, or be necessary to implement, subsections (a-c), counties with fewer residents than the district population shall be assigned completely to a single district whenever possible, and to at most two districts.7
5. Except where doing so would conflict with, or be necessary to implement, subsections (a-d), a district shall be convex, which is to say that the convex hull of a district shall not contain the centroid of a county, or portion of a county, that is assigned to a different district.8
6. Except where doing so would conflict with, or be necessary, to implement, subsections (a-e), a district shall be compact, which is to say that the convex hull of a district shall not have an aspect ratio exceeding the twice the largest aspect ratio of the convex hulls of all counties that are within or partially within that district.9
7. Except where doing so would conflict with, or be necessary, to implement, subsections (a-f), district boundaries within a county shall be drawn to minimize splitting of minor civil divisions, such as cities, towns, townships, villages, and reservations.10
In each reapportionment subsequent to the enactment of this act, one third of seats within a state’s delegation, rounded down, shall be designated as proportional seats voted for indirectly, with representatives who shall be considered to represent the state as a whole. A state shall therefore have a number of apportioned districts equal to two-thirds of the number of apportioned seats, rounded up to the next whole number.11
1. Candidates voted for directly (“principal candidates”) shall be grouped into sets of candidates based on voluntary mutual political affiliations (“parties”) upon or before the date of their qualification as principal candidates.12
2. These parties shall, as necessary, include new parties defined by a single principal candidate who is unaffiliated with any previously recognized political party, but has otherwise qualified as a candidate for election as determined by state law.13
3. All such parties shall have the ability to designate a list of additional reserve candidates who shall be considered as candidates to represent voters of that party from the state at large. Each reserve candidate shall be a citizen eligible for service within the House of Representatives and shall not be a reserve or principal candidate appointed by any other party.14
4. Ballots cast in elections for the House of Representatives shall include both an expression of preference between principal candidates for that district and an expression of preference between all parties competing within the same state. These expressions of preference shall be combined from votes or points into individual and party numeric totals (“support”) as otherwise prescribed by law, with the caveat that for determining party support, a voter shall be able to mark support for a single party, and no voter may contribute a greater total number of votes or points of support across all marked parties than a voter who has marked a simple preference for a singular party.15
  1. For each district, the principal candidate with the largest individual support in that district shall be assigned a seat.16
  2. Each party’s remaining quota shall be calculated as Q = D*(P/T)-A, and seat ratio shall be calculated as S = D*(P/T)/A, where D is the total size of the delegation, P is the total support for the party, T is the total support for all parties, and A is the number of seats assigned to that party.
  3. For each unassigned proportional seat, if there exist parties with Q ≥ 1, A=0, and at least one remaining candidate, the party with the highest remaining quota Q among zero seat parties shall be assigned that seat. Otherwise, if there exist parties with Q ≥ 1, S > 1, and at least one remaining candidate, such party with highest seat ratio S shall be assigned that seat. Otherwise, the party with the largest remaining quota Q and at least one remaining candidate shall be assigned that seat.17
  4. The at-large representatives shall be assigned to proportional seats in a manner designed to geographically approximate the distribution of voters not otherwise represented by a candidate of their chosen party. In particular, for each assigned proportional seat, if there are districts with candidates from that party and no winning candidates from that party, that seat shall go to the principal candidate who earned highest individual support within those districts, or, in the absence of any such principal candidates, a reserve candidate who resides within the district with the highest party support among such districts. If no such districts remain, the next priority shall be similarly among candidates who ran for or are resident within districts with one winning candidate from the party, then as a third priority, candidates within districts with two winning candidates, et cetera.18
  5. Whenever the procedure in (iv) shall identify a set of multiple candidates or shall be unable to operate, the seat shall respectively be assigned at random to one eligible candidate from the set of multple candidates or from the whole list of eligible candidates of that party.
In each reapportionment subsequent to the enactment of this act, the number of members of the House of Representatives shall be the cube root of the total population counted within the immediately preceding decennial census, rounded to the nearest integer.19
1. Section (a) of the Permanent Apportionment Act of 1929 (2 U.S.C. § 2a) is amended by striking “the then existing number of Representatives” and inserting “the cube root of the the whole number of persons in the United States, excluding Indians not taxed” in its place.20
2. If the next general election shall take place more than a year from the enactment of this act, and no regular apportionment is scheduled to take place prior to that election, a special reapportionment shall be initiated upon enactment using the most recent decennial census data, in accordance with the procedures established under 2 U.S.C. § 2a as amended herein. If no regular reapportionment is scheduled to take place prior to the general election after the next general election, a special reapportionment shall be initiated after the next general election using the most recent decennial census data, in accordance with the procedures established under 2 U.S.C. § 2a as amended herein.21

Discussion

My proposal attacks both means and motives by objectively making it more difficult to gerrymander and by adding compensatory seats that will get handed out to parties that get locked out of too many individual seats. It eases the way by increasing the size of the House, which has three key benefits.

Preserve the degree to which individual members of Congress are responsive to their constituents.
Allow most incumbents to be re-elected by winning or performing as a strong second in districts that largely overlap their current districts.22
Ensures that most states are affected by proportional rules.23

Keeping the House at or near the cube root of the population is historically ordinary. From the 8th to 87th Congresses, the House was apportioned to no less than 80% (typically ~90%) of the cube root of the most recent decennial census population.24 I’ll discuss a few questions I expect to hear often, and then step through a few example states to talk about how difficult it would be to gerrymander hard enough to overwhelm proportional representation.

Why not require independent redistricting commissions?

This proposal neither bans nor requires independent commissions, which I favor as one of many available tools to handle gerrymandering. It is my hope that state legislatures will more readily delegate the power to draw district lines to a nonpartisan appointed commission or to an algorithmic process once the stakes are low enough to make partisan gerrymandering a low-return process.

Unfortunately, while independent commissions have resulted in significantly better maps in the states that have used them, the set of criteria used to create independent commissions and the rules under which they operate is a matter of political contention, and even under some of the best-designed commission rules, the parties can and will work to try to subvert the independence of commissions.25

Why use proportional representation at all?

It is possible to minimize the gerrymandering problem by instituting fully algorithmic redistricting, and Congress does have the ability to impose algorithmic redistricting on the states the same way that they have the ability to ban multi-member districts, at-large slates, et cetera. I am working on a redistricting algorithm that I think is objectively fair.

However, absent any change in the current level of polarization, partisan support for adopting an algorithm will fall along the lines of whichever party’s experts think they will get a small advantage using that algorithm. The only way to get proportional representation reliably is to impose some measure of it directly.

Why not a national closed-list system?

Many political scientists love national closed-list systems based on success stories in other countries. American voters, on the other hand, have a long history of distrust for insiders making decisions in smoke-filled rooms, and House delegations are assigned by state on a basic constitutional level; American federalism makes national proportional representation unworkable. Americans want to vote for candidates on an individual basis, and most members of Congress have an individual approval rating among their constituents that is significantly higher than their national party.26

Shifting to a closed-list system even at the state level would be a major change that would result in many individually popular incumbents being ousted in favor of faceless party insiders. While I understand that many experts loudly favor closed-list systems, this proposal is designed to be both agreeable to Americans at large and minimally disruptive outside of what is necessary to address gerrymandering. In this case, that means that most elected candidates will be ones that have their name on the ballot within a district similar in size to current districts.

Is one-third of seats enough?

Mathematically speaking, reserving only one-third of seats isn’t enough to ensure that we can compensate for the very worst gerrymanders. Many proportional representation systems include “overhang” seats to restore proportionality. However, within the American context, reserving one-third of seats is sufficient in most cases after the application of very basic restrictions on gerrymandering. The precision of an American proportional representation system is limited to plus or minus five seats.

Example state: Utah

Utah with Salt Lake City cracked four ways in a gerrymander (left) or mostly captured in a central Democratic-leaning district (right).

Utah’s state legislature enacted a partisan gerrymander in a flagrant violation of the democratic process after voters decided to hand districting over to an independent commission. Utah is one of the easy illustrations of cracking. In the cube root proportional allocation, it would have five districts (instead of the current four districts), with one district contained entirely within Salt Lake County, one mostly contained within Salt Lake County, and two compensatory seats, similar to the current map rather than the partisan gerrymander.

Democrats are unlikely to lose a compact Salt Lake County district while earning enough votes to merit a third Congressional seat out of a seven-seat Utahn delegation; conversely, they would struggle to win more than two districts in a year where their performance merited two seats.

Example state: California

The lightly (and mostly racially, per a “communities of interest” standard) gerrymandered map drawn by the independent commission (left), and the egregious partisan gerrymander installed by Prop 50 (right).

The Proposition 50 gerrymander, replacing the fairer maps drawn by one of the nation’s most famous independent redistricting commissions, has fifty-two districts, seventeen of which are fully contained within a county and thirty-five of which cross county lines. If the current map had to follow the county division rules above, the proportions would be reversed: Only fifteen districts would be able to cross county lines, with thirty-seven being contained entirely within populous counties.

In a typical California election year, Congressional Republicans earn a bit under 40% of the vote. Under my proposal, California would have fifty-six districts and twenty-seven compensatory seats, meaning California Republicans would only need to win six districts to obtain a proportionate delegation in a typical election. Minimal county division would place seven districts entirely within Republican-leaning counties (Riverside, San Bernardino, Kern), five entirely within narrow battleground counties (Fresno, Orange), and ensure that smaller counties make up most of the population of most of the sixteen multi-county districts.

Example state: North Carolina

The nine(!) congressional districts used to crack & pack Wake, Mecklenburg, and Guilford into a 10-3 map.

North Carolina, which has a long history of egregious partisan gerrymandering, would have fifteen districts and seven compensatory seats, one entirely within and one mostly within each of Wake and Mecklenburg counties. A 12-10 or 11-11 delegation would be proportionate for any recent North Carolina election in this battleground state.

The current district map, with fourteen districts, is a hard gerrymander that cracks the state’s three largest counties, Mecklenburg (D+16), Wake (D+12), and Guilford (D+11), across three districts each. Without those extra county divisions, it is difficult to draw fewer than two Democratic-leaning districts in the Charlotte area and fewer than three Democratic-leaning districts across the Triangle and Triad, and five seats are all that are needed to support a proportionate 12-10 break.27

Closing comments

I welcome questions, objections, and suggested revisions in the comments, and will update both the model act above and the Q&A section to reflect useful feedback you provide. If you think that what you’ve seen is good enough with no notes required, please send it to your representative, your senator, and any politically engaged citizens you know.

Thanks for reading MathIntee Political Post! This post is public so feel free to share it.

For the 2020 census, this would have been 692 seats. This keeps districts about the same size in the transition from single-member districts to mixed-member proportional. This has theoretical justification as well as historical precedent; the House was set to within 20% of the cube root of the census population in every apportionment act passed from 1800 to 1950.

This constraint by itself eliminates most of the really egregious gerrymanders, though by itself, it tends to produce maps with a Republican lean due to Democrats engaging in more “natural packing” in high-population counties.

A 2-seat state using proportional rules would effectively just have multiple at-large candidates. A 2-seat state using compact districts will generally have one small district around its major population center and then a large “hinterland” district. In most small states, this would add up to one competitive district and one safe district, though which is which would vary (c.f. Maine, Montana). A county with population almost exactly matching the district population quota, to within measurement error, is unlikely but could happen.

This is a separate clause from (d) in order to make it a higher priority, since these counties are already being split into multiple districts and gerrymandering within a county is not as easily controlled.

This rules out “corner-connected” counties and requires partial counties to use most of the adjacent boundary. A typical county-county boundary is around 25% of the perimeter of the smaller county; less than 10% means the counties really are not connected very closely. I have referred to the convex hull of a county here because some counties, notably coastal counties and counties with water boundaries, may have unusually large or hard-to-calculate boundary lengths.

This clause means that if you needed just a little tiny piece of an adjacent county to balance population, that’s fine, as long as it’s not a deep piece.

The most complicated water adjacency situation is New York with its heavily inhabited islands, most notably Long Island and Manhattan, but it also comes up to a lesser degree for North Carolina’s Outer Banks, Virginia’s Eastern Shore, Hawaii, and various other islands. Defining the water-crossing adjacency as one unique pair of points provides a significant objective constraint.

Since you can add whole counties to a district one at a time, the number of small counties that are divided should generally be n-1-k, where n is the total number of districts and k is the number of districts completely contained in large counties. In rare cases, whole county populations could line up neatly to produce the district population quota to within measurement precision.

This is a convexity requirement that bans bypassing a county by curving the district lines around it. The expected mode of violation is that a high-population county could be surrounded by two lower-density districts, one of which contains the centroid of at least one interior district within the populous county.

This is a weak compactness requirement that functions to rule out long “snakes” of low-population counties. The expected mode of violation is via contiguity and long skinny portions of states, such as the Maryland panhandle and Oklahoma panhandle.

There are twenty states with town or township subdivisions.

We’re in America. Weird independent candidates and third parties are supposed to crop up every so often.

This is not completely free of potential problems.

Among other things, bringing in “reserve” candidates at this stage helps address the California-style “top two” primary, which can leave a D-D or R-R contest, or reintroduction of “approval” style voting.

For individual candidates, this is loose enough to allow for not only the plurality votes used in most state elections or ranked choice voting as used in Maine and Alaska, but a wide variety of other voting methods as they may be adopted, such as STAR voting or a Borda count. The additional mathematical constraint means the party vote is essentially forced to be plurality vote or a cumulative vote, however; allocating proportional to party support is the one area in which a plurality vote is fairly functional.

I.e., we’re starting off by doing exactly the same thing we’re already doing for at least 2/3 of the seats.

This prioritizes “shut out” parties first, then underrepresented major parties, then finally minor parties.

Essentially, what’s happening here is this selects the at-large candidates who are geographically proximate to the largest pools of unrepresented voters.

The cube root of the 2020 Census population would round to 692.

This puts the cube root law in place on a permanent basis.

In practical terms, this means that it takes effect as immediately as possible, with the allowance of one calendar year for states to redraw districts before holding primaries and an exemption saying that a special reapportionment doesn’t have to take place if a regular reapportionment under the amended process would take care of it anyway. Given that multiple states demonstrated the ability to redraw districts as little as six months before the 2026 election, a full calendar year is sufficient; the new apportionment by state can be calculated nearly instantly based on existing apportionment formulas. Modern computerized tools make drawing districts fairly easy.

In order to be implemented, this act would need support from members of Congress who would expect to face re-election under the new rules. This is designed to be a sensible incremental reform.

Under 2020 Census apportionments for 692 seats, eight states; Maine would have been the lowest-population state with three seats, and would probably retain almost exactly the same two-district map under this proposal.

It was apportioned at under 0.8x the cube root only once before the 1960 census, in the 1790 census. The Civil War arguably produced sub-0.8x representation, but Congress had not passed a new apportionment act. The apportionments of the 1820 and 1830 Censuses slightly exceeded the cube root of the census population, and the trend had brought the House very near to the full cube root of the population again a century later when a political crisis related to the 1920 census hit, resulting in the size of the House being essentially frozen for the next century.

I would consider California an example of this. The commission maps have been very close to fair from an efficiency gap perspective, and meaningfully prevented the most egregious partisan gerrymandering of the type put into place with Proposition 50, but it’s also clear that California Democrats worked very hard to try to subvert the second independent commission to produce more favorable outcomes, mostly by using the “communities of interest” criterion, and it is accordingly very difficult to get Republicans to support the California style commission.

The fact that our parties are incredibly unpopular is a major problem in and of itself, but also a political reality.

Wake and Mecklenburg both essentially account for two districts each. Guilford isn’t large enough to trigger (1)(b) and force an interior district, but if divided into two pieces, one of the two Guilford districts would have to extend east into super-Democratic Orange or west into significantly-Democratic Forsyth. (To the north, Stokes, Rockingham, and Caswell combined don’t add up to enough population to fill out a partial-Guilford district, and convexity will bring in Orange or Forsyth at that point.) I would even say it seems pretty difficult to build a safe 10R-5D gerrymander with 15 districts; IMO, the most natural map will have several battleground seats with a 8R-7D median performance, reflecting a small Republican advantage in concentration driven mostly by Mecklenburg, Durham, and Orange.

The Cato Institute is wrong about immigrant crime

Tomas McIntee — Tue, 31 Mar 2026 21:48:22 GMT

The Cato Institute was founded in 1977 to support Reagan’s version of the Republican Party. It was and remains a political advocacy organization rather than a producer of high-quality research, and continues to side with the interests of capitalists as a global class using arguments designed to appeal to the American bourgeoisie while nakedly opposing the interests of the American proletariat.

Subscribe now

As the Republican Party has now returned full-circle to being a worker’s party with nativist and protectionist tendencies, the Cato Institute is no longer loyally Republican. However, their status as part of a rump “Reagan Republican” group that has not yet fully assimilated into the Democratic coalition has not changed the quality of their scholarship.

The Cato Institute begins with their desired conclusions and engineers evidence to support those conclusions. The most recent Cato study to cross my feed is a methodological disaster zone on crime and immigration. It fails, unfortunately, to account for either recidivism or misclassification; a simple back-of-the-envelope adjustment for this shows their results are easily nullified or reversed by reasonable corrections for known bias effects.

This can be debunked in a short and simple manner.

Immigration and crime

The Cato Institute has consistently advocated for open borders since 1981. Not content with claiming that immigrants are just as good as Americans, they have generally tried very hard to claim that immigrants are better than Americans, which makes it much easier to argue that we should have much more immigration. Their latest work is a claim that immigrants are less criminal than Americans.

The first problem with this table is that criminal immigrants are supposed to be deported after their first offense, and that persons with a criminal history are not supposed to be admitted in the first place. Fewer than 31% of federal prisoners have no prior convictions.1 If immigrants are uniformly deported permanently following a first conviction, the underlying criminal risk profile is almost identical.2

Correcting Cato’s toplines for two known effects produces much more nativist numbers. The first step is probably an overcorrection, especially since previously-deported illegal immigrant criminals do return for second incarcerations.

Another major problem is misreporting of status. Criminals are presumed native-born citizens unless directly recorded otherwise, and an incarcerated immigrant whose immigration status escaped being recorded upon incarceration has a strong incentive not to bring that status to official attention.3 While it is impossible to directly measure this skew, a reasonable proxy is the low quality of ethnicity data for arrests and convictions. One-quarter to one-third of Hispanic criminals are demographically misclassified as white by default in public records.4

The above estimates are not, by any means, precise figures. There are substantial additional problems and complications in trying to calculate illegal immigrant crime rates, including underreporting of crimes with illegal immigrant victims and witnesses, divisor difficulties (no good estimate of total illegal population), and sanctuary cities and states choosing not to incarcerate criminals vulnerable to deportation. Nor are they useful figures.

C Smoe rime rates in Denmark by national origin, including both the top 10 nations of origin by total population and top 10 for total number of criminals. (Male-only to avoid sex-selection skews.) Data from StatBank.

The immigrant population is not some uniform “non-American” blob. It is a sample from a large and diverse global population, with cultural and socioeconomic factors that vary significantly from country to country. Denmark is one example of a country for which national origins data for criminals is publicly available, and their data shows significant heterogeneity among the immigrant population, including between neighboring countries with related populations but distinct cultural factors.5

The Cato Institute’s analysis in favor of open borders is very obviously flawed on multiple levels; the only thing that it shows is that the Cato Institute’s analyses should be disregarded as a general rule. In the real world, human migration is a complex phenomenon involving a combination of positive and negative effects, and every nation has an obligation to its citizens to try to minimize the negative effects as well as maximize the positive effects.

Thanks for reading MathIntee Political Post! This post is public so feel free to share it.

From here, federal prisoners have a 31% chance of being in the lowest criminal history category, which includes a modest number of prior offenses but mostly consists of first-time offenders.

31% of 1,195 per 100,000 would be about 370 per 100,000 vs the actual of 303 per 100,000 for legal immigrants and a whopping 674 per 100,000 for illegal immigrants. For Hispanic immigrants in particular, the predicted value would be around 390 per 100,000, significantly lower than the actual value of 486 per 100,000.

Sanctuary jurisdictions also appear to undercharge immigrant criminals in order to avoid causing deportations.

See mainstream media coverage here, but also the tweet below.

@iamyesyouareno Texas alone has over 12,000 Hispanics listed as \"Non-Hispanic White\"...","username":"UnbiasedCrime","name":"Unbiased Crime Report","profile_image_url":"https://pbs.substack.com/profile_images/1672724900635901953/aBaw8ori_normal.jpg","date":"2024-02-28T10:13:09.000Z","photos":[],"quoted_tweet":{"full_text":"Here is the tally @TxDPS has corrected so far...\n\nHispanics booked as Non-Hispanic White: 0 of 12,011\nBlacks booked as White: 18 of 106\nAsians booked as White: 0 of 38\nWhites booked as Black: 0 of 1","username":"UnbiasedCrime","name":"Unbiased Crime Report","profile_image_url":"https://pbs.substack.com/profile_images/1672724900635901953/aBaw8ori_normal.jpg"},"reply_count":10,"retweet_count":48,"like_count":268,"impression_count":20161,"expanded_url":null,"video_url":null,"video_preview_media_key":null,"belowTheFold":true}" data-component-name="Twitter2ToDOM">

Within Denmark, people from Iraq have twice the crime rate of people from Iran. People from Pakistan have eight times the crime rate of people from India. People from Romania have a little over twice the crime rate of people from Ukraine. People from Greece (not pictured on the graph) have far lower rates of criminality than people from Turkey. These differences are quite significant, and are related to the circumstances of migration as well as differences in crime rates within source countries.

Gun Control & the Glockalypse

Tomas McIntee — Sun, 15 Feb 2026 19:11:50 GMT

I originally intended to write an article rehashing the great revolver debate, but there are plenty of articles that cover it out there:1 Semiautomatics broadly offer superior firepower, but revolvers are broadly more reliable.2 That article would have gone up on Tomas’s Miscellany, an article that only incidentally noted the political and legal situation.

However, while I could find good coverage mechanics of guns online, anything related to politics or law is usually swamped by a sea of poorly-presented arguments of dubious merit, and there have been some interesting new developments in gun control issues that deserve discussion.

So, instead of writing a more technically-oriented article, I’m going to talk here on the MathIntee Political Post about the Glockalypse and the past and likely future of gun control in the United States, only incidentally noting that I think semiautomatics are the wrong default handgun3 and that focusing on volume of fire over accuracy has negative consequences.4

The Glockalypse

From July 1st, 2026 forward, the sale of certain new guns in California will be forbidden.5 The law targets semiautomatic guns that can be easily converted to fully automatic operation, singling out a mechanism used by Glock pistols and numerous generic Glock clones. The problem: Glock switches, invented three decades ago, are small, simple, easy to manufacture,6 and easy to install… and transform a semiautomatic pistol into a small machine gun.7

Glock switches as advertised on a website selling them. These devices are designated as “machine guns” in law. As a side note, I will focus on actual English language definitions in this instead of picking and choosing from a variety of legal fictions.

This is part of a broader problem: Semiautomatic weapons are generally one mechanically simple change away from fully automatic action, which were invented earlier.8 They were developed from automatic weapons by interrupting an otherwise fully automatic action. Since almost all of the mechanical energy of the reloading action comes from the chemical energy of recoil, continuous fire is a short mechanical transformation away.9

It has proven nearly impossible to prevent criminals from getting their hands on Glock switches due to the fact that it is a simple inert part; a switch-enabled Glock with a high-capacity magazine10 can provide modern Chicago gangsters with a similar volume of fire to the classic Thompson submachine gun, used by Chicago gangsters in the previous century in the high-profile incidents that inspired the original 1934 National Firearms Act.11

Shortly after California passed its new law, Glock withdrew its entire current line of pistols and replaced it with a new line, the Series V, with the intention of ending the Glock switch phenomenon.12

Politics & law

The reason that it’s difficult to have any kind of discussion about guns is that there are two very loud and very vocal extremes well-represented: People who want no gun control at all and people who want guns banned. The legal reality is inconvenient for both sides of the debate:

Neither state nor federal law can ban guns on a sweeping basis.
States and the federal government can regulate sale, transfer, and ownership of firearms on rational bases.
Volume of fire (both capacity and rate of fire) is an established rational basis for regulation of firearms.

This is similar to how abortion rights were treated by the courts from 1973-2022,13 but as it involves much more direct readings of constitutional law and many more precedents in the history of American law and preceding English common law, it is a state of legal affairs that is likely to continue indefinitely.

The famous Tommy gun whose use by Prohibition-era gangsters inspired an entire wave of gun control laws. (Source.)

One major accepted rational basis for gun regulation, first reaching national prominence with the 1934 National Firearms Act, is limiting access to weapons that can easily produce high volumes of fired bullets. Michigan imposed a sixteen-round capacity limit on firearms at the state level in 1929; currently, ten-round capacity limits are common.14

The mathematics that supports the core rational basis

Based on the available data on civilian self-defense, each bullet fired has a significant probability of being the last bullet needed to either incapacitate or deter a criminal attacker.15 This means that the marginal defensive utility of additional rounds of ammunition decreases very substantially with each shot. Outside of a statistical tail consisting mainly of panicked “mag dumps,” it is very rare that self-defense shootings involve more than five rounds.16 In fact, a significant fraction of defensive gun use involves deterrence without shots fired, though a gun known to be empty is not very effective as a deterrent.

The basic structure of the imbalance between utility for defensive use (socially positive) and offensive effects of criminal use (socially negative). Crossover point will vary based on assumptions used.

Criminal use of weapons involves a much higher volume of fire. Five rounds is the average number of shots fired at police officers in a gunfight with a criminal.17 Drive-by shootings and mass shootings 18 are on an entirely different order of magnitude when it comes to volume of fire; gang violence sends stray bullets through residential neighborhoods on an unfortunately frequent basis.

For legitimate defensive use of firearms, higher capacity has rapidly diminishing returns. For reckless or criminal use of firearms, the negative social consequences are nearly linear with the number of shots fired, and the number of shots fired tracks much more closely with the the capacity of the gun and its effective rate of fire. The combination of these two facts gives a rational regulatory basis for restricting weapons with a high effective rate of fire, such as automatic weapons and high-capacity weapons.

The end game

There will not be a Dobbs decision for Second Amendment rights, because the right to bear arms is, unlike a right to abortion, written directly into a constitutional amendment.19 Much like abortion regulation during the Roe period from 1973-2022, the ability of states to regulate firearms for capacity and rate of fire depends entirely on the interaction between pragmatic and technological realities and the vagaries of the court system. Arbitrary regulation is likely to get struck down.

Targeted technical regulation, like California’s Glock ban, will tend to survive if it has a rational basis. I expect California’s Glock ban to survive. And those regulations will build up over time. This is how anti-abortion states made it very difficult for residents to have abortions, and the only reason that it has not gone further in anti-gun states is that gun control advocates have not previously taken a sufficiently technical approach to targeting specific mechanisms.

Since most existing semiautomatic weapons can be modified for automatic operation and can accept high-capacity magazines, most existing semiautomatic weapons can be banned. While it is not likely that any state will ever be able to effectively ban all semiautomatic weapons, it is likely that any particular semiautomatic weapon will be eventually banned in at least one state.

In a short nutshell, it’s simpler to teach how to operate a revolver safely, easier to check that the weapon is unloaded and safe, easier to do dry training with, it encourages more thoughtful and accurate shooting, has greater tolerance for variation in powder charges, is more likely to be fully operational after it and its ammunition have been stored for an extended period of time. I would emphasize the diminishing return of additional shots on safety (most self-defense incidents involve 3 or fewer shots), which is discussed some in this article, and also that the majority of gun owners, gun users, CCW permit holders, et cetera can be expected to go through periods in which they let ammunition and/or weapons sit for extended periods of time in less than ideal climate-controlled circumstances.

With high-quality ammunition fresh from the store and regular maintenance, semiautomatics are now very reliable mechanically. However, operator error, maintenance failures, and out of spec or old ammunition can still easily lead to jams.

This doesn’t mean that a semiautomatic handgun doesn’t have its plus points over a revolver… but on the balance, high volume of fire is only rarely a significant net positive. For most people, reliability, robustness, simplicity of operation, and being consistently legal in most states are all more important factors.

One of the most interesting points broadly in favor of revolvers being considered comes not from civilian usage, but from police department training and usage patterns. While revolvers are not appreciably more or less accurate on a physical basis, limits on volume of fire seem to encourage greater emphasis on accuracy.

Broadly, most of the old gun grognards seem to agree that the adoption of semiautomatics by police departments took place around the same time as a broad decline in accuracy standards for police departments. On a purely theoretical basis, it’s reasonable to guess that since volume of fire and accuracy are partially fungible, declines in accuracy standards would create pressure for the switch and the switch would in turn reduce the incentive to maintain high accuracy standards.

Like most of the gun control laws actually passed, this is not quite a ban on the possession or transfer of guns and has no impact on the used firearm market, so the actual impact of the “Glock ban” is limited.

They can even be 3D-printed.

“Machine gun” in the legal sense of a fully automatic weapon. “Machine pistol” would be the more accurate term.

In fact, “semiautomatic” pistols were originally called “automatic pistols.” The Maxim machine gun, the earliest modern automatic machine gun to see battle, was invented in 1884 and in service from 1886 to the present; the first semiautomatic pistols showed up about a decade later.

Bump stocks are one of the more indirect methods of doing this. They allow the recoil to move the main part of the gun, therefore moving the trigger mechanism back and forth against a depressed trigger finger to provide a rate of fire of about ten rounds per second.

Both the switch and the high-capacity magazine are illegal in Illinois, to be clear.

With the weight distribution, optional barrel shroud, and foregrip, the Tommy as desined is a more effective and durable way to deploy that volume of fire than a Glock with a switch and high-capacity magazine, but volume of fire is the central problematic element.

Since the Series V still uses a cruciform trigger bar, it seems unlikely that the Series V will be available in California after the ban date. It is also likely, given design lags, that the Series V was under development prior to the passage of the ban; however, it seems clear it was intended to stymie the switch phenomenon, which has been a source of bad press for Glock.

To spell out the similarity more explicitly, the courts had held that:

Neither states nor the federal government could ban abortion on a sweeping basis.
States and the federal government can regulate abortions on rational bases.
Fetal viability is a rational basis for regulation, along with various health and safety provisions.

In general, neither pro-abortion nor anti-abortion advocates embrace philosophical positions that justified the legal nuance of fetal viability that was at the center of the distinction Roe v. Wade drew, which is similar to both pro-gun and anti-gun advocates today.

New York passed a law with a seven-round loading limit, while retaining a ten-round magazine capacity limit; however, the seven-round limited was struck down on the basis that the seven-round loading limit was arbitrary. The ten-round capacity limit remains in place.

See here for discussion of hits. In general, criminals retreating or fleeing is more likely than incapacitation as an outcome. Additionally, the defender may become incapacitated or lose control of their firearm prior to firing all available rounds of ammunition.

A notable example of an outlier: Kyle Rittenhouse fired a total of eight rounds in one of the most prominent recent cases of shootings ruled to be self-defense, involving an unusually large number of aggressors. The widespread expectation by trainers is the “Rule of Threes” (three yards or less, three seconds or less, three or fewer shots), and the typical number of defensive shots appears to be two.

See here.

The highly-visible mass shootings involving semiautomatic rifles are rare, but not measurably rarer than defensive shootings in which a high-capacity magazine proved beneficial.

The logic involved in the Roe v. Wade decision (which you can read here) in finding a limited right to abortion up to the point of viability can be generously described as “complex,” and it is not surprising that the interpretation neither existed for the first century of the 14th Amendment nor that it was eventually overruled. I still suspect that the 9th Amendment would have been a much more straightforward method of discovering a pre-existing unenumerated right to bodily autonomy

End political violence.

Tomas McIntee — Sat, 17 Jan 2026 19:12:22 GMT

I learned about the assassination of Charlie Kirk while sitting in a quiet little coffee shop in San Francisco, pecking away at my laptop as I tried to focus on work. A barista was chatting with a friend; both were young women, by all appearances perfectly ordinary.

A man had been killed, one barista explained; she thought it was a good thing, he deserved it. One wasn’t sure who Charlie Kirk was; the other clarified that he the guy who went around college campuses arguing for right-wing stuff, which meant he was asking for it as far as she was concerned. I was a little unsettled to hear the cheerful reaction to an assassination and the swift justification thereof; I also knew next to nothing about Charlie Kirk.

The core problem: A sharply rising minority of young left-wing Americans celebrate political violence when it targets their enemies. This poll was conducted shortly after the assassination of Charlie Kirk; the baristas I overheard were not outliers for their combination of demographics, location, and political affiliations, though unusual within the American population as a whole.

I know a lot more about Charlie Kirk now than I did when he was alive, and I’ve seen a lot more happen since then. This article will discuss political violence in liberal democracies from Charlie Hebdo to Charlie Kirk and beyond, a paradox of tolerance widely attributed to Karl Popper, and the urgent problem of rising political violence in the United States, three distinct political topics that are nevertheless closely related. I encourage you to share this article with friends and tense acquaintances across the political spectrum.

There are a lot of people who do not want to believe that there is a real and growing problem with political violence in the United States, or that the problem is growing on the left. It is urgent that Blue Team liberals and moderates take steps to identify and isolate the violent part of the left wing from the larger Blue Team political tent and work to de-radicalize the left.

Right-wing political violence

For a lot of readers who are liberal, left-wing, or just generally have Blue Team loyalties, the first response to the above assertion is to claim, without good evidence, that right-wing political violence is a greater problem.1 Right-wing political violence exists, but it is neither seeing the same growth nor is it being actively ignored.2 Currently, the primary right-wing danger to the democratic process does not lie in political violence.3

The aftermath of the Oklahoma City Bombing in 1995, perpetrated on the 220th anniversary of the Battle of Lexington (traditional date of Patriots’ Day) by a man whose politics were broadly right-wing and appeared to have a right-wing political motive.

For example, anti-abortion violence targeting clinics and providers has continued after the Supreme Court struck down Roe v. Wade (1973) in Dobbs v. Jackson (2022); however, it has declined. The volume and deadliness of anti-abortion violence peaked sometime in the 1990s.4

Generally, right-wing political violence has been a higher priority for domestic law enforcement than left-wing political violence, being more vigorously investigated and prosecuted by state and federal authorities for most of the last several decades.5 The most alarming instance of right-wing political violence since 1995,6 the January 6th storming of the Capitol, was curbed by law enforcement and punished with prosecutions.7 Now is not the time to shout tu quoque in order to avoid cleaning one’s own house.8

The reality of the problem

Charlie Kirk’s assassin9 was a left-wing political independent motivated by left-wing politics.10 The act was abetted and celebrated by people on the left.11 This information was actively concealed by Blue Team media, with 24% of respondents believing that the assassin was a Republican and 17% believing that his motives were right-wing.12 After Jimmy Kimmel spread the lie that the shooter was aligned with rival right-wing groups, Blue Team media rallied in defense of Kimmel even as evidence that the assassin was left-wing continued to mount.13

Scene of the attack. (Source.)

The shooting of Charlie Kirk was not an isolated event.14 A significant amount of recent political violence in the US can be reasonably be described left-wing based on either the motives or the affiliations of the attackers.15 Other cases had less clear motives and assassins with unclear affiliation, such as the high-profile attempt to assassinate Donald Trump that left him with a bloody ear.16 Comparatively few have been right-wing.17

These attacks have targeted not only Republicans, but other perceived enemies of left-wing causes on the basis of that emnity, including certain moderate Democrats,18 Jews,19 Christians,20 health care executives,21 whites,22 ABC affiliate stations,23 and ICE.24 Many of these violent acts have been celebrated by left-wing individuals on and off social media.25

Left-wing political violence is a growing trend, and a lot of moderates in the Blue Team tent are in denial about this problem, partly due to rising partisanship within the media ecosystem.26 While there has been significant left-wing political violence in the US in the past, what we’re seeing right now is a recent trend that emerged in the 21st century.27

Je suis Charlie

There are real and related global trends in political violence. The slogan Je suis Charlie dates from ten years ago, referring to Charlie Hebdo, a French magazine that is frequently irreverent and satirical. When that irreverence and satire has been turned on Muslim subjects through its cartoons, the response has included lawsuits (2006), firebombings (2011), shootings (2015), social media censorship (2020), and stabbings (2020); essentially, all possible venues have been pursued, both legal and illegal, to censor the magazine and punish those involved with its production.28

Some of the controversial content published by Charlie Hebdo.

Globally, Islamists are responsible for a large amount of political violence, and particularly political violence intended to suppress controversial speech. However, America did not have a significant domestic population of Islamists until changes in immigration patterns during the Bush II administration, and the population remains a small minority.29 Islamist political violence used to be detached from mainstream American politics.

From foreign phenomenon to the American mainstream

Islamists have fallen into company with the American left wing. This is a very new phenomenon. Before the 21st century, Muslims were simply not a significant presence in mainstream American politics as far as either major party was concerned.30 In the Trump era, Islamists may be found integrated with democratic socialists in political and protest networks.

A pro-Palestinian encampment on a US university campus. The Palestinian cause has been a major focal point of cooperation. (Source.)

The alignment of Islamist organizations with the American left wing has been broadly associated with a rise in tolerance for political violence31 as well as being directly linked to a number of specific acts of political violence.32 This relationship isn’t a simple unidirectional cause and effect; both trends reinforce each other and share common root causes.

The postmodern left is censorious, equating speech with violence.33 There have been plenty within the postmodern left who responded to the terrorist attacks on Charlie Hebdo with criticism of Charlie Hebdo. The roots of this particular postmodern left alignment trace back to the 1960s, explained and exhibited quite clearly through examining the work and life of Frantz Fanon.

This is the cover of the most visible version of Frantz Fanon’s most popular book in the US, in my experience; it’s the version I own and can easily be identified from a distance. Image here is fair use without permission.

For most mainstream Democrats of the 1990s, Frantz Fanon was a fringe radical they hadn’t heard of; for the growing postmodern left that worked its way into the mainstream in the 2010s, his work and ideas were gospel.34 Unfortunately, those ideas include a direct endorsement of revolutionary violence.35 In addition to the embrace of explicitly violent ideology, there are four key ideological factors that have contributed to the growth in left-wing violence:

Moving away from universal principles, especially the principle of free speech.36
Failure of the liberal political establishment to recognize and address the illiberality of the postmodern left.37
Treating disagreeable speech as a form of violence.38
Finding hateful subtext in statements by political opponents.39

I didn’t invent this meme, but I find it has unfortunate resonance. Treating speech as violence while discounting or even celebrating actual violence is a commonly noted feature of the postmodern left.

The last two points combined are particularly relevant to the Charlie Kirk assassination and the response to it.40 Within the “blue bubble,” all Red Team views are construed as highly radical and hateful, with moderate positions concealing radical subtext that is revealed or constructed by Blue Team media elites, through close reading and Freudian psychoanalysis if necessary.41

Factually, most of Charlie Kirk’s political opinions, both on the issues and, more importantly,42 on the subject of how issues should be decided, were not especially radical. His opinions would have been considered largely ordinary within the Republican Party from the 1980s to the present, and most would not have been remarkable in a moderate or conservative Democrat of the 1990s. However, they did not need to be radical to be construed as radical and hateful by observers within the blue bubble.43 The rise of partisan media has produced a significant problem with strategic partisan misinformation.44

Charlie Kirk’s main project was a devotion to civil debate, dialogue, and discourse. His harshest critics within Red Team’s media ecosystem were the true violent radicals.45 His assassination has made the future ascendance of right-wing political violence more likely, rather than less, both because he was a likely future leader himself and because he was in a position to significantly influence other likely future leaders like J. D. Vance.

Where do we go next?

The problem we face right now as a society is one that Karl Popper described in 1945. Our survival as a liberal democracy that relies on discourse and debate requires that we identify and suppress those who advocate for and engage in political violence. This is widely described as the “paradox of tolerance.” In order to maintain a tolerant society, we cannot tolerate the substitution of violence for persuasion.46 Democracy needs a loyal opposition, one that accepts a common democratic framework and is willing to accept losing elections.47

“But we should claim the right to suppress them if necessary even by force; for it may easily turn out that they are not prepared to meet us on the level of rational argument, but begin by denouncing all argument; they may forbid their followers to listen to rational argument, because it is deceptive, and teach them to answer arguments by the use of their fists or pistols.” - Karl Popper, 1945

Popper was right. This is why January 6th, 2021 needed to be taken seriously.48 This is why the Boogaloo Boys should be (and were) prosecuted. And, in the wake of the Charlie Kirk assassination and other recent acts of left-wing political violence, it is why the Democratic Party and the associated Blue Team media ecosystem needs to kick out the radicals who look fondly upon political violence.49

In the past, the Democratic Party was much better at pushing the radicals to the margins and included a more significant number of moderates. The growing Horseshoe Caucus, on the other hand, is full of radicals who are actively trying to eliminate the moderates from the party. The minority of House Democrats who voted against condemning Kirk’s assassination represent the leading edge of a troublesome trend.

Political violence is not going to preserve our democracy.

This is, generally, an example of tu quoque argument. Deflecting from the issue doesn’t address the problem.

Blue Team and Red Team partisans will both claim that their side has always been less violent. Both sides can present convincing and earnest research supporting that claim. The methodology of most of this research is suspect; for example, many such studies do not include the George Floyd riots or the January 6th insurrection as political violence at all, in spite of both being highly significant episodes of political violence. Others are skewed by singular high-mortality events, such as the 9/11 attacks and the Oklahoma City bombing. However, while comparative baseline levels of political violence in the 1980s-2020s varies significantly, the trends generally show stagnant right-wing violence and increasing left-wing violence.

I would say that the dangers are executive overreach and undermining the legitimacy of the electoral process, both of which can still be addressed both in the court of law and the court of public opinion. The Trump administration controls neither the court system nor the electoral process; indeed, it has thus far been unable to deport one particular high-profile individual who unambiguously lacks the legal right to reside in the United States because it has by and large (if sometimes belatedly) obeyed the orders of the court system.

This is most trivially true with the inclusion of the 1995 Oklahoma City Bombing and measuring by total deaths, but it’s also true counting incidents broadly. A small amount persists. The most notable recent act of political violence by a right-wing figure was the assassination of Melissa Hortman, which was carried out by a person known to be strongly opposed to abortion. While the assassin was incontestably right-wing, whether or not the violence was motivated by his right-wing perspective is contested; he claimed to have been acting on the orders of Governor Tim Walz and may have been motivated by delusion.

Here, I’m speaking of the 1980s onward; the place where there is some room for debate on this topic is Islamist political violence, which has sometimes been a high priority for federal law enforcement. Some people would classify this as right-wing (as it is religious conservatism) and others left-wing; for example, there was a string of three prominent pro-Palestinian acts of political violence in the spring of 2025, with the first two being perpetrated by ordinary domestic left-wing activists and the third being a flamethrower attack by an Egyptian pro-Palestinian attacker on June 1st.

The fact that foreign Islamists and domestic left-wing activists may be conducting similarly-targeted political violence for similar reasons poses some difficulty for trying to classify political violence by categories.

Whether the successful Oklahoma City Bombing of 1995 was more or less alarming than the failed Counting Day Putsch of 2021 can be debated at length.

The Trump pardon of the rioters was a dangerous precedent. For better or worse, the punishments meted out prior to the mass pardon were substantial. Some may debate whether or not I’m right to label the event as the Counting Day Putsch and mark it as the most significant episode of right-wing political violence within the US in the last twenty years, but while the number of deaths due to the violence was small (1 rioter) and the disruption failed, this marked the first forcible disruption of the process of counting electoral votes in the entire history of the United States.

Educational footnote: The tu quoque fallacy, named for a Latin phrase meaning “you too,” is also known as “whataboutism.” Rhetorical deflection to the sins of the “other side” is a very common phenomenon; unfortunately, many loyal partisans are hypocrites, and we can’t really deal with wrongdoing without building conditional or transient alliances between those who are loyalty to principles over partisanship and self-interested hypocrites who would not support prosecution of their own side.

As assassins are sometimes motivated by a desire for notoriety, I have left the perpetrators in this article nameless and faceless, even though I discuss their motives and affiliations, and would encourage others to do so when practical.

The evidence for this is quite conclusive: Texts indicating he was motivated by the claim that Kirk was spreading hate, his family testifying about the political division that had grown between them, and his lover’s trans activist posts on Reddit. The bullet casing inscriptions alone would have been considered conclusive with a right-wing attacker. It is also unambiguous that the attacker, while politically left-wing and with motives legibly informed by “Blue Team” descriptions of Charlie Kirk, was not a Democrat.

While it is not clear if the assassin had assistance prior to the act, it is known that an unrelated left-wing individual stepped up to falsely confess to the assassin in an attempt to help the actual assassin escape. Which, frankly, is wildly alarming. Celebrations on social media were widely documented and in some cases subsequently deleted to avoid a wave of right-wing cancellation attempts.

Jimmy Kimmel’s prominent claim that the attacker was “MAGA” (Trumpist) was not grounded in any evidence at all. The attacker also was not a “Groyper” by any accounts.

Texts to the assassin’s roommate, testimony of his family, and other evidence emerged later as events unfolded.

Most notably, The Atlantic recently ran an article from experts who believe that 2025 is the first year that left-wing terrorism has overtaken right-wing terrorism in the United States, tracing the start of the trend of rising left-wing violence to Donald Trump’s election in 2016.

The distinction between these two things is difficult but real. Even in the era of high political polarization, many passionate people do not subscribe to a fully left-wing or right-wing package of beliefs. The health care CEO assassin is a good example of someone who was not really right-wing or left-wing overall (we have his Reddit posts to analyze) but had a clear legibly left-wing motive laid out in his manifesto (he talked about health care companies being “parasites,” “too powerful,” and “greedy and corrupt,” citing left-wing activists Michael Moore and Elisabeth Rosenthal).

The fact that the assassination attempt was lauded by many on the left, however, is a problem in and of itself regardless of the motives of the shooter.

The assassination of Melissa Hortman was carried out by a right-wing individual who declared his motives were to enable the election of Governor Tim Walz as the next Democratic senator for Minnesota. The general consensus is that he was acting on the basis of insanity, with a left-wing minority insisting it was anti-abortion violence, a right-wing minority insisting it was part of a pro-immigrant agenda, and no significant factions celebrating the assassination.

On April 13th, 2025, a left-wing American activist attempted to murder the sleeping moderate Democratic governor of Pennsylvania and his family by setting the governor’s mansion on fire in the name of Palestine. If he had succeeded, this would have been the first assassination of a sitting governor in more than a century.

On May 21st, a left-wing American activist carried out a mass shooting that targeted at least four people and resulting in the deaths of two Israeli embassy staffers outside the Capital Jewish Museum during an event organized by the American Jewish Committee, ironically an event intended to build a coalition to help Gazans. By all appearances, the target selection was based simply on identification of the targets as presumptively Jewish.

Such as the Covenant shooting in Minnesota, see here.

As mentioned in Note 13, while the assassin who killed Brian Thompson was not clearly identifiable as left-wing in his everyday life, his stated motivations (health care costs are driven by greed and corruption), beliefs, and the sources of those beliefs (Elisabeth Rosenthal, Michael Moore) all were left-wing, as were those who applauded his act.

Notably, a stated motivation in a 2023 Tennessee shooting.

On September 19th, a left-wing retired lawyer who had spent twenty years working for Blue Team political groups, including as the legislative advocacy director for California’s teachers’ union, fired a gun at Sacramento’s ABC affiliate station in a drive-by shooting.

On July 4th, eleven people wearing tactical gear and armed with guns attacked an ICE facility in Texas in an unambiguous paramilitary assault by an organized left-wing group. On September 24th, a lone gunman used a bolt-action rifle to fire from the rooftop of an immigration lawyer’s office into an ICE facility and a windowless ICE van. In the four months since I drafted this article, there have been numerous additional incidents of violence.

I personally have seen shrines to the healthcare CEO assassin at essentially random convention vendor booths.

Both on the left and right wing, news has become heavily filtered by a desire to make one’s own team look good and the other team look bad. Left-wing political violence looks bad for Blue Team, so it’s mainly discussed on right-wing media.

Especially notable: Weather Underground, responsible for multiple bombings and arson attacks in the late 1960s and early 1970s.

I can’t find a more comprehensive and concise review than “Controversies” section of the Wikipedia article.

Nation of Islam would generally be classified as a black nationalist group, with distinct goals and methods even if they are nominally Muslim.

The most prominent Muslims in American politics in the 20th century were Malcolm X and Louis Farrakhan, who were both well outside of the mainstream. Keith Ellison, the first Muslim member of Congress, began his political career in 2002 and was elected to Congress in 2006.

As with many social phenomena, it can be difficult to distinguish cause and effect. Higher tolerance for political violence has enabled greater support for Islamic terrorists within the American left; conversely, embracing groups and activists broadly aligned with Hamas, the Muslim Brotherhood, and other Islamist groups pushes the left into

I already mentioned several examples, but pro-Palestinian political violence also occurred during street protests, such as the murder of Paul Kessler in Thousand Oaks, California in 2023, less than a month after Hamas’s October 7th attacks.

Indeed, in general, postmodern people tend to treat words as having magical importance, bringing literary symbolism to the lens through which they view reality. Since postmodernism comes from a generally literary rather than scientific background, this is not surprising.

Fanon is among the highest tiers of the most-cited writers on Google Scholar in spite of his relatively short career. The short of it is that Frantz Fanon took ideas from Marx and Freud and applied them to race and colonialism in order to justify terrorism, rape, ethnic cleansing, and genocide by Algerians. In analysis that has been thoroughly disproven by the subsquent course of history, he believed that revolutionary violence was both more effective at gaining independence and resulted in a better long-term outcome for the successor state.

It’s worth noting that Frantz Fanon was not speaking purely academically; he fought for the FLN in the field, wrote propaganda for them, and assisted with their military strategy, and was therefore directly responsible for some measure of political violence by an Islamist terrorist organization.

Or, to use a postmodern turn of phrase, I might say something like “rejection of the normative Enlightenment meta-narrative.” The idea that everyone should operate by the same universal principles and standards and has a certain set of basic fixed rights is at the core of liberalism.

A few specific individuals have done so; however, these individual warnings have rarely translated into effective action, and many formerly liberal institutions are now aligned fully with the postmodern left.

Burning a Quran or an American flag, or calling someone by a racial slur, is not an act of violence. Classifying disagreeable speech as an act of violence creates a justification for using violence in response to disagreeable speech. This attitude is not unique to postmodernism; instead, what’s closer to unique is the classification of disagreeable speech and insults to a group or cause as fundamentally nonviolent, which is a cornerstone of the globally and historically distinctive kind of philosophical liberalism that social scientists sometimes describe as WEIRD.

If someone says “silence is violence,” this is what they’re doing.

It is possible that Kirk was targeted because he was moderate and therefore effective at persuading undecided voters; however, the assassin characterized Kirk as “full of hate and spreading hate,” and most likely believed Kirk to be radical.

Misrepresentation of opponents as radical is not limited to Blue Team, though I would say from experience (and polling) the Blue Team epistemic bubble is more thoroughly closed than the Red Team epistemic bubble, mainly because a greater share of traditional media and entertainment elites are aligned with Blue Team and partly because Red Team has more ideological diversity. (See, in particular, the Obama-Trump voters.)

The distinction between radical political opinions and radical opinions about how politics should be practiced is subtle but important. For example, the Amish widely hold genuinely radical beliefs, but this is politically nonproblematic because they do not attempt to use violent force to impose their radical lifestyle on others. Conversely, the idea that the private health insurance industry is broken and needs to be reformed is mainstream, but assassinating healthcare CEOs in order to try to frighten health insurance companies into self-reform is a radical and unacceptable political tactic.

For example, Karen Attiah misquoted Charlie Kirk, and then numerous Blue Team media figures doubled down, claiming the misquotation was word-for-word accurate.

Broadly, both sides benefit in the near term by painting the opposition as more extreme than they really are. In the long term,

This includes Groypers and Boogaloo Boys; mainstream Blue Team media went as far as to highlight the fact that there was a a disagreement between Groypers and Charlie Kirk, but many tended to omit the fact that the disagreement centered on Kirk’s insistence that democratic debate was at the center of meaningful political change.

Popper’s paradox is often summarized as saying that a tolerant society cannot accept intolerance, and then misused as a justification for political violence against persons deemed “intolerant,” such as racists or sexists. Flatly, it’s clear that liberal democracy can function and progress in the presence of widespread bigotry; the risk seen with the rise of Hitler, Lenin, and other dictatorial regimes is the use of political violence to suppress opposition.

Since every party is eventually out of the majority position, this means every party has to be willing to accept that an agenda contrary to their desires will be implemented from time to time.

Trump’s “sore loser” problem when losing the 2020 election was troublesome for the same reasons that partisans refusing to accept the lost election as legitimate was troublesome in 2000, 2004, 2008, 2012, 2016, and 2024. However, right now, the Republican Party is playing within the rules compatible with Popper’s standards of tolerance; they are trying to win elections rather than eliminate their political opposition by force of arms, and they did indeed win democratic elections in 2024.

Broadly, the Trump administration can and has been checked both by the courts and Congress; at present, the fact that the administration has thus far been unable to permanently deport an illegal immigrant who had never received a visa on essentially technical and procedural grounds should be taken as firm evidence that the court system remains effective as a check on abuse of executive power. The democratic process was largely functional in the 2024 elections, and the 2026 midterms are likely to remain as democratically functional as the 2018 and 2022 midterms were, provided that there is not a widespread outbreak of political violence.

Given that the radicals are the ones pressing purity tests on candidates, I firmly believe that Democrats will find that they gain a bigger tent of voters and politicians if they kick out the radicals.

Proposition 50 was a mistake

Tomas McIntee — Sun, 21 Dec 2025 19:03:20 GMT

Remixed from 1 2.

The United States regularly redraws electoral districts every ten years in response to the decennial census. Rarely, a state will redraw its Congressional districts in the middle of a decade. Unfortunately, both political parties have engaged in a race to the bottom.

As soon as the 2020 census concluded, Democrats clamored for as much gerrymandering as possible, saying that “unilateral disarmament” required them to gerrymander as hard as possible. After Trump was elected in 2024 with an exceptionally narrow House majority, he called for Republicans to gerrymander as hard as possible. In both cases, this was a tactically sharp yet strategically boneheaded move.

Donald Trump was poised to make himself the new face of gerrymandering by demanding that Republican state legislatures draw him new mid-cycle maps. Fortunately for Trump, Gavin Newsom wanted to share the spotlight.

Once the Texas state legislature responded to the call, Gavin Newsom, sitting governor of California and the most likely apparent next Democratic president,1 decided to spearhead a highly public effort to gerrymander California’s congressional districts for the 2026, 2028, and 2030 elections. This was, again, a sharp tactical move… but strategically inept.

This, unfortunately, was a bad idea: Republicans had more room for mid-decade gerrymandering available to them than Democrats, and the Trumpmander campaign faced opposition that Gavin Newsom has helped quell. Proposition 50 was a small tactical-scale political victory with larger but harder to see strategic-scale political costs, with Gavin Newsom abandoning the high ground to join Trump in the swamp.

While it’s impossible to test a counterfactual scenario directly, it’s likely that the entire Proposition 50 campaign, from start to finish, has decreased Democrats’ chances of taking back the House in the 2026 midterms.

Some background

Both parties have a long history of drawing gerrymandered maps; California has a long history of particularly egregious gerrymandering intended to shield incumbents from competitive elections. In spite of Democratic opposition, California voters chose to enact major electoral reforms in 2010, eliminating partisan primaries and creating a nonpartisan electoral commission responsible for redrawing electoral districts.2

The 2010 electoral cycle was a wave election for Republicans thanks to their strength among high-propensity voters in a low-turnout midterm cycle, giving them control of most state governments; they also engaged in a boldly overt campaign of gerrymandering nationally. This contributed to the perception that the party was antidemocratic and elitist,3 fueling a popular vote loss in 2012. However, the gerrymandering was effective enough that Republicans held an unearned House majority in the 113th Congress - though not in the other four elections of that redistricting cycle.4

The wild 2012 election. (Source.) Darker shades indicate pickups.

Some of the hard gerrymanders of 2012 broke into dummymanders in the 2018 wave election following the first stages of the Trump realignment; however, even after the Democrats won and kept a House majority during the final two elections of the census cycle, loyal Democrats continued to decry gerrymandering as a Republican offense against democracy.

During the redistricting cycle following the 2020 census and Trump’s loss to Biden, Republicans were divided between pro-Trump and anti-Trump camps and had recently lost many of their experienced operatives and planners. They had neither the ability nor the inclination to gerrymander as effectively as they had in 2010.

In 2021, a surge of thinkpieces like this one blasted “unilateral disarmament” by Democrats.

Democrats were in a more organized state and engaged in a boldly overt campaign of gerrymandering nationally, with Blue Team media elites writing op-eds declaring that gerrymandering was both good and necessary. This contributed to the perception that the party was antidemocratic and elitist,5 helping fuel a popular vote loss in 2022.6 On the face of it, this was similar to Republicans in 2012.

However, due to the fact that Democrats were not able to perform partisan gerrymanders in New York7 and California, as well as the natural packing linked to minority-majority districts (“VRA districts”), the Democratic gerrymandering campaign of 2020 was less effective than the Republican gerrymandering campaign of 2010. Republicans won the popular vote and a majority of the seats in 2022 and 2024; not as many as expected from their margin, but still a working majority… as they deserved in exchange for winning more votes overall.8

The VRA district problem

Until quite recently, it was standard orthodoxy among the Democratic Party that the Voting Rights Act of 1965 required a generous number of majority-minority districts. “VRA seats” are generally safely Democratic, creating strong representation for minority factions within the Democratic caucus. Without significant gerrymandering efforts, minority-majority districts tend to “pack” Democratic voters inefficiently, to the overall advantage of Republicans; however, the presence of these districts also prevents extreme partisan gerrymanders by both parties.

Minority-majority (including majority-minority) districts of the 119th Congress. Source. Many are geographically small, so they don’t show up well on the large map.

The Supreme Court is imminently likely to change the rules in the forthcoming Louisiana v. Calais case; however, the Proposition 50 map was drawn under the old rules. Even friendly legal experts say the drawing process was “race conscious,” which was fine under the old rules, but could be more strictly forbidden by the time the Proposition 50 challenge inevitably makes its way to the Supreme Court.

It is possible that Louisiana v. Calais will open up a fresh round of gerrymanders in several states. Louisiana itself seeks to eliminate its lone majority-minority district in favor of an extreme partisan 6-0 gerrymander.

Continued tit-for-tat

As demonstrated by Indiana Republicans’ refusal to pass a Trumpmander, Donald Trump is not able to simply wave his hand and erase all Democratic seats in all Republican-controlled states. The Proposition 50 campaign, highly public and very visible nationally, helped provide Trump with political capital for additional Trumpmandering.

The 2024 map was favorable to Democrats; when the Proposition 50 campaign began, Republicans had a greater potential for seat gains via midcycle gerrymandering than Democrats, making escalation a risky choice. The proposition was not guaranteed to pass, faces a serious chance of being blocked in court, and illustrates sharply that Democrats’ most likely 2028 standard-bearer is willing to get down in the mud with Trump.9

There is no scenario in which both parties cheating more leads to a healthier democracy. The correct response to the Trumpmander was to draw contrast, seize the higher ground, and deprive him of the political capital needed to pressure multiple state legislatures.

Collateral damage & the bigger picture

California’s gerrymander-shielded delegation can be expected to be less responsive, less representative, and more corrupt.10 Additionally, because the effort requires a public campaign with substantial public advertising, and Gavin Newsom is nationally prominent as one of the leading contenders for the 2028 Democratic nomination, the Proposition 50 gerrymandering effort in California involved a greater public commitment to gerrymandering by Democrats than the corresponding Republican efforts in Texas, Florida, Indiana, and Missouri combined.

By supporting Proposition 50, Democrats visibly abandoned the high ground, setting their sights lower than a breaking wave. This is high ground that Democrats desperately need. Democrats are at the nadir of their popularity. I mean nadir very seriously: The modern Democratic Party is a continuous organization that dates back to the Jacksonian branch of Thomas Jefferson’s Democratic-Republican Party, and to the best of my knowledge, it has never been less popular than it has become during the Trump era.

No amount of gerrymandering will allow Democrats to win the Senate or the presidency; and it is unlikely that retaliatory gerrymandering will win Democrats the House unless they abandon minority-majority districts.11 Perhaps worse, by practicing justification of gerrymanders, reform becomes an even more distant prospect.12

Based on current betting market odds on the 2028 nomination and general election, plus the fact that nobody really knows what the field will look like in 2032 or later.

The most notable supporter of this effort was the very heterodox Republican governor of the state, Arnold Schwarzenegger.

Because it was.

In 2014 and 2016, Republicans won a majority fair and square, and in 2018 and 2020, Democrats did well enough to break through the gerrymander. As with the Electoral College, unearned majorities are rare.

Because it was.

This loss was widely and erroneously framed as a “win” within the blue bubble.

Because one party usually has an advantage with high-propensity voters and the other with low-propensity voters, it’s common for American elections to be countercyclical, with the president’s party losing ground in low-turnout midterms (e.g., Republicans in 2010 and 2014). Blue Team media outlets largely framed the loss as being less than typical midterm losses and therefore actually a win, which helped Democrats delude themselves into thinking that they were still on track to win a majority in the 2024 elections.

The fact of the matter is that in spite of a documented advantage in high-propensity voters following the Trump realignment, Democrats lost the House in 2022.

Partisan gerrymander struck down by the courts and replaced with an independently drawn map. This was replaced by a Democratic gerrymander for the 2024 cycle.

Mid-cycle partisan gerrymanders in New York and North Carolina between the 2022 and 2024 elections roughly balanced each other.

This is probably good for Gavin Newsom’s chances of nomination and bad for his conditional probability of winning the general election given nomination.

Gerrymandering has these known effects.

If the Supreme Court rules as expected, then Democrats will have to disown majority-minority district drawing practices in the next cycle, at least publicly; but based on my current understanding of internal party politics, Democrats will not voluntarily abandon the practice of deliberately drawing majority-majority districts.

The Rucho decision was that Congress has the power to ban political gerrymandering and has thus far failed to exercise it. Neither party has chosen to do so, although both parties have had opportunities to do so.

How should we end gerrymandering?

Tomas McIntee — Mon, 15 Sep 2025 17:13:37 GMT

Founding Father (and Massachusetts governor) Elbridge Gerry and the eponymous salamander-like “Gerry-mander” district.

Gerrymandering, the practice of drawing the boundaries of electoral districts in order to gain a political advantage, is enjoying one of its perennial moments in the public eye. As I have discussed previously, this is a problem that the Founding Fathers did not anticipate, and which poses multiple difficulties for democracy.1

What I want to do today is to take the first steps in building a fair and transparent algorithmic approach to drawing districts. The goal is to provide a model for avoiding gerrymandering.2 This is slightly complicated by the contradictory legal standards surrounding racial gerrymandering, but I strongly believe that a facially neutral system is the best basis.3

Measuring district quality

The goal of the algorithm will be to approximate the following attributes:

Maximize contiguity between county and district boundaries.4
Minimize division of population centers.5
Maximize overall compactness of the district map.6

As a basic fact of geometry, given n districts, it should be possible to draw a map contains no more than n-k-1 instances of a partial district within a county, where k is the number of districts that can be drawn inside of the boundaries of a heavily populated county. Any iterative approach that draws one district at a time using whole counties and adds at most one partial county with each new district drawn will include at most n-1 cases where a partial district is within a county.

Since county lines generally do not vary much over time, are usually not redrawn arbitrarily, and usually serve as boundaries for election administration, county lines serve as a reasonable basis for an algorithmic approach. I will formalize an algorithm using county lines and the US Census’s subdivisions (census tracts, et cetera).

My previous approach

In order to provide a point of contrast during my discussion of the Rucho case, I gave an example of drawing districts for North Carolina that used the following steps:

Identify the n largest population counties as possible centers for districts.
From this pool, eliminate counties that are small enough to be potentially combined into a single district and adjacent to each other. This gave me the kernels for eight districts.
Identify natural regions of geography (Appalachia, coastline).
Fill in the gaps, trying not to split cities or counties unnecessarily.

This is not a very easy to formalize approach. The results were a perfectly fair map for use in comparisons to the real-life maps, but while the process was fairly mechanical, it is still difficult to formalize and may not work as well for different physical distributions of populations.

Outline of a proposed algorithm

So, I would propose the following process:

Identify all counties with populations large enough to contain at least one district and reserve within-county districts to those counties.
Select a new county to start a district with.
Add whole counties by shared boundary metrics until the population quota is met, then add a last partial county.
Repeat steps 2-3 until no unassigned counties remain.
Divide counties based on the same principle of minimum boundary length using census tracts.
Repeat step 4 with census blocks within census tracts.

States are divided into counties or county-like objects; counties are divided into census tracts; census tracts are divided into census blocks.7 Census data is freely available, and the census subdivisions are directly linked to the population data that serves as the basis for apportionment.

Properties and puzzlements

How do we pick a new county in Step 2? One choice would be to pick up with the partial county that we left off with in the previous cycle, which minimizes division of counties. Another choice is to look out… as in literally outward.

The key thing that made filling in my original NC map straightforward was the fact that after reserving districts within Wake and Mecklenburg that would not cross county lines, I essentially worked my way from the outside of the state inward. This was caused by my focus on physical geography, looking at the mountains and the coastline. Once I had finished the outermost four districts and two inner metropolitan districts, filling in gaps was straightforward.

The VRA problem

The Voting Rights Act has been interpreted as both forbidding and requiring a certain amount of racial gerrymandering in order to create majority-minority districts. Majority-minority districts also tend to be naturally “packed” districts that distort the overall partisan fairness of the map, an effect that was crucial in creating the Republican gerrymander advantage of the 2010 cycle that led to a rare example in 2012 of a party winning a majority of House seats without a plurality of the vote.

This is a complicated issue that cannot be solved easily or fairly by algorithm because there is no consistent objective standard that can resolve the difference between “good” and “bad” racial gerrymandering under the historical standards used to apply the VRA.

Not entirely coincidentally, it is thought likely that the Supreme Court will decide entirely against both “positive” and “negative” racial gerrymandering before the next redistricting cycle. This, in turn, means that algorithmic redistricting is likely to become practical even in states where natural compact districts will not include many majority-minority districts.

Call for commentary

So, what do you think? Please leave feedback and suggestions in the comments, and I will incorporate those in my discussion in my next article.

The majority in Rucho v. Common Cause claimed otherwise, but factually, this is based on a misunderstanding of the historical record. In modern terminology, the term “gerrymandering” is not used to refer to malapportionment, which was the primary method of drawing districts for political advantage prior to the writing of the US Constitution.

Broadly, districts that are drawn compactly and algorithmically without political considerations cannot be “gerrymandered” by any reasonable definition of the term. The problem is advance coordination.

Essentially, the Voting Rights Act has been read as both requiring and prohibiting racial gerrymandering in the form of minority-majority districts. There is no robust method for achieving proportionality of representation by a politically distinct minority in a single member district system that does not involve deliberate gerrymandering. The methods I outline above will lead to the election of minority candidates to the degree that those minorities control geographically coherent and compact communities and to the degree that districts are of sensible size.

I.e., minimize the number and severity of cases where a partial district may be found within a county.

Gerrymanders frequently take one of two form factors: A long “snake” or “tentacle” format, such as the original salamander-like “Gerry-mander,” or a radial pattern around a population center. The radial pattern emerges as an attempt to either dilute or amplify the power of a major population center.

Defining “compactness” is not trivial.

Counties or county-like subdivisions include parishes, independent cities, etc.

The Biden blame game

Tomas McIntee — Sat, 19 Jul 2025 18:06:56 GMT

Jake Tapper’s recent book, “Original Sin,” has triggered a wave of recriminations across the partisan divide. What the book has brought into public view: The public was systematically deceived about the degree to which Joe Biden’s health—including energy levels and mental acuity—declined over the course of his term.

Share MathIntee Political Post

The problem with analysis of the book is that we are now operating in a low-trust environment. Every party involved—including Biden’s inner circle, the rest of the administration, and the members of the news media with inside access—has incentives to lie in order to try to shift the blame elsewhere. Almost all of them also have loyalty to one or another specific political agenda.

Tapper himself was involved with publicly downplaying Biden’s age during the campaign. Many other media figures involved were his friends and colleagues.

Would Biden have done worse?

The first falsehood in Tapper’s book is one that I’ve discussed repeatedly: The claim that all evidence suggested that Biden would have had a spectacular loss “far worse than that suffered by his replacement as Democratic nominee, Kamala Harris.” It’s possible that Tapper believes this false claim.

However, it is very obvious in the publicly available polling data that the candidate swap was accompanied by an unprecedented increase in Trump’s support. The claim that Trump would have gained similarly against Biden is speculative; it’s not supported by “all evidence.”

These graphs show Trump’s polling with the candidate switch marked. 538’s polling average as well as independent analysis of their collection of polling data paints a very clear picture of the effect of switching candidates: It sharply increased support for Trump to a level not precedented by pre-swap data.

Kamala Harris was a bad candidate, and there’s strong evidence that she was held back by her individual weaknesses as a candidate:

It was clear all the way back in 2019 from head-to-head polling, approval polling, and past electoral performance that Kamala Harris was a weak candidate.
In fact, in 2019, I projected that Kamala Harris would run about 1.9% behind the median Democratic candidate nationally, revising that hypothetical projection downward to 2.5% in an updated analysis in January of 2020.
As I detailed after the election, she ran badly:
- 1.5% behind the median Democrat in NC’s eleven statewide elections, meaning 2.5% worse than Biden or Clinton.
- 2% behind the median Democratic senatorial candidate, meaning 0.9% worse than Biden.
- 2.9% ahead of the median Democratic gubernatorial candidate, meaning 3% worse than Biden.
Issue polling showed that voters believed that Kamala Harris still stood behind positions she took while campaigning for president in 2019.1

All quantitative evidence available points to a cynical conclusion: If Democratic elites had continued to stand by Biden after the debate, minimizing it as an outlier in his public appearances, the race would likely have remained largely static, leading to a closer race.2 This evidence includes the “instant reaction” polling and focus groups used to gauge voters’ response to the debate.3

Instead, Democratic media elites spearheaded a campaign to take down Biden, relying heavily on the lie (repeated by Tapper in his book) that Biden had promised to step aside after one term rather than loudly and repeatedly denying that he was considering such, and ignoring his better appearances, including a press conference in which he energetically lectured hostile reporters on foreign policy.

Following the switch, however, there was a strong immediate and negative response by swing voters, followed by a continued negative trend as less-informed swing voters learned more about Kamala Harris.4 Everyone knew that Biden was old; not everyone knew what Kamala Harris stood for, but they would rapidly find out.5

I’m posting these graphs of Trump’s polling numbers twice because it shows something extremely important about the polling from 2024.

As I have said repeatedly, Kamala Harris didn’t just fall out of a coconut tree; she was burdened by what had been, and voters held her to account based on her record. In 2020, the main line of attack by Red Team against Biden had been his advanced age and the specter of a possible Kamala Harris presidency. In 2024, the main line of attack by Red Team was educating voters in detail about Kamala Harris’s record.

I suspect that many voters, including most swing voters, would have preferred a Biden presidency wherein Biden was largely a figurehead who had delegated his authority to the chosen members of his staff to a Kamala Harris presidency wherein Kamala Harris was actively making decisions. The few voters who would have preferred a Kamala Harris presidency were rarely swing voters, but rather reliable base voters who would have held their noses and voted for Biden hoping for his death or resignation.6

As I have also pointed out numerous times, voters have been willing to vote for aging politicians in poor health even in circumstances where a reasonable observer would expect that would mean that their staff or vice president would end up playing a large role in the running of their administration.7

Biden’s condition

What I know about Biden’s condition right now is almost the same as before the release of Tapper’s book:8

Biden was having disturbing “senior moments” even in 2020.
His public appearances were limited and carefully managed.
He nevertheless still had many public appearances where he sounded perfectly coherent.

While I am skeptical that Tapper’s book is accurate, there is one datum that Tapper provides that I think I can trust, because it is a claim that could easily be falsified:

Biden’s meetings with important political insiders were limited.

For a president to have limited public appearances is ordinary. For a president to have limited direct contact with important legislators of both parties is extraordinary. For a president to have extremely limited contact with his own cabinet secretaries, as alleged by Tapper, is astonishing. It was already known that full cabinet meetings were rare and scripted, but that could have been because of their size and semi-public nature.

Executive branch org chart. Cabinet secretaries report directly to POTUS, or in exceptional circumstances temporarily to VPOTUS.

It is not quite unprecedented. Edith Wilson, the wife of Woodrow Wilson, essentially interposed herself between the outside world and her invalid husband for the last seventeen months of his term. This was controversial.9 It also predated the 25th Amendment, which clarified that a president incapable of performing his duties should be relieved by his vice president, and was the result of a truly alarming situation. The fact that Biden was secluded by his inner circle convinces me that the core of Tapper’s accusation is correct:

Biden was no longer able to reliably fulfill the duties of the presidency, and instead of the presidency devolving upon the vice president, his presidency was largely controlled by a circle of unappointed insiders.

This isn’t something I wanted to believe. Biden’s public appearances were compatible with a man who was an old gaffe machine, past his prime but still competent to delegate the small things and decide the big things. Instead, it appears that an inner circle did step in between Biden and the outside world, managing both the information getting to Biden and the decisions made by Biden. While Biden may have made a significant number of decisions, especially in his more lucid public appearances, the intermediate authority of Biden’s inner circle usurped the federal chain of command in much the same manner as Edith Wilson.

Who deserves blame?

When I look at what has happened with the Democratic Party during the Trump era, I see a long string of mistakes with a common origin: Rapid polarization, disconnection, and radicalization of elites through social media. However, what we’re dealing with right now is a little bit more specific, and the problems start in 2019 with the race for Democratic nominee.

The cold underlying reality is that Biden’s original candidacy violated the William Henry Harrison rule: Do not nominate candidates who are 68+ years old. Most violations of the William Henry Harrison rule are recent.10 While it is true that some individuals remain capable past the age of 68, a president who is first elected at age 68 is very likely to have reduced capacity within the next eight years.

William Henry Harrison’s presidency consisted of a month-long illness. His death precipitated a significant political crisis.

The Biden team is partly to blame for violating this norm. Political and media elites were complicit in a series of failures that resulted in Biden winning the nomination anyway. Interestingly, hhe primary process itself was not the problem. Even the candidate selection process for the official DNC debates was not much of a problem.11

Instead, the problem was that Democratic political elites remained relatively quiet, while Democratic media elites were extremely vocal about who should and should not be nominated.12 As I wrote in an article for Towards Data Science, there was clear bias in both volume and tone within media coverage of the chaotic Democratic primary field.13 This can be summarized as follows:

Bias in volume of coverage in favor of Acela Corridor candidates.
Bias in tone of coverage against straight white males.
Bias in both tone and volume of coverage in favor of progressives and against anti-establishment candidates.14

Various Democratic media elites tried very hard to push forward less-electable candidates, including Elizabeth Warren, Kamala Harris, Pete Buttigieg, Cory Booker, and Julian Castro.15 People within the “blue bubble” wanted a progressive pro-establishment candidate who ticked at least one demographic quota box. The problem was that none of these candidates was popular outside of the blue bubble. In fact, retrospectively, it’s likely that any of the establishment progressive candidates would have lost to Trump.16 The strongest candidates were the ones sharing the moderate lane with Biden. The candidates generally checked few diversity checkboxes and usually came from outside the Acela Corridor.17

Four candidates who could have replaced Biden in the moderate lane if they had been supported by Democratic media elites instead of actively attacked by Democratic media elites: Amy Klobuchar (Minnesota), Beto O’Rourke (Texas), Michael Bennet (Colorado), and Steve Bullock (Montana).

However, this “lane” of candidates was starved for attention. Most Democratic voters wanted a credible moderate, and Democratic media elites systematically tried to exclude credible moderates like Amy Klobuchar and Beto O’Rourke who could potentially have displaced Biden in his lane. Similarly, they failed to elevate anti-establishment alternatives to Bernie Sanders.18

In an extraordinarily crowded contest, media elites did have enough influence to choose which of the more obscure candidates was elevated to contend against Biden and Sanders. Other than the late decision of the New York Times to halfway endorse Amy Klobuchar (an endorsement split with Elizabeth Warren), Democratic media elites were united in trying to keep electable alternatives to Biden off the Democratic ballot, and were largely successful in suppressing the visibility of those candidates within a crowded field.

Finally, once political elites and Democratic voters rallied around Biden as the only candidate who seemed likely to defeat both Sanders and Trump, the decision to nominate Kamala Harris as vice president instead of a credible moderate successor to Biden compounded the problem. With Kamala Harris as his second, Biden could not step aside without handing the presidency to a deeply and persistently unpopular progressive of questionable competence.

The original sin was not the decision of Biden to run for re-election. It was not even the decision of Team Biden to isolate the president from his own cabinet. Instead, the largest and most significant mistakes happened in the lead-up to the 2020 election:

Joe Biden (or his team) choosing to run in 2019.
Democratic media elites choosing to embrace unelectable candidates in the lead-up to the primaries in 2019.
Joe Biden (or his team) selecting Kamala Harris as a running mate.

These came before Joe Biden (or his team) decided to run for a second term in office. And there are three clear generalizeable lessons in this.

The William Henry Harrison rule should be respected.
Media elites should not be trying to play kingmaker.
Vice presidents should be picked based on merit, not for reasons related to symbolism or internal party politics.

Still, while not the original sin of this debacle, the fact that the Biden administration appears to have been controlled by an unelected and unappointed self-described “Politburo” is disturbing. This should be investigated seriously; if true, there should be serious consequences.

See here for some of the most detailed issue polling. Blueprint labels these beliefs “misperception,” but there is no question that Harris loudly and vocally backed most of these positions in 2019 even if they were not included in her official 2024 platform.

There is no guarantee that Biden would have won, but given that Trump only won by a margin of 1.5% of the popular vote, it can’t be ruled out as a possibility. As I pointed out, voters have elected dead candidates before.

Focus groups watching the debate showed little to no shift in voter opinions, and neither did before-and-after polling. Some even showed a shift in favor of Biden. As far as providing fodder for advertising, many short marketable clips showing Biden visibly confused or stumbling over his words were available prior to the debate.

Red Team aired numerous highly effective attack ads that consisted largely of accurately describing Kamala Harris’s statements and actions.

Kamala Harris did not have a high profile as vice president, and she had dropped out of the 2020 nomination contest before the primaries started.

One of the sharpest examples of this is that based on message testing, most of Red Team’s favorite attacks against Biden that emphasized the possibility that Biden would resign or die, or that Kamala Harris would be running the presidency behind the scenes.

Most dramatically in 2000 in Missouri when voters chose to elect a dead man to the Senate, but see also Strom Thurmond and Dianne Feinstein, Ronald Reagan in 1984, and William Henry Harrison in 1840; Harrison’s ill health and Reagan’s early signs of Alzheimer’s were both contemporary campaign issues that seemed to have zero impact on public support. Note also the 2020 campaign, where Biden was old enough that his death or decline in office would be considered a real risk by any sensible voter.

Interestingly, a large number of examples of Biden’s cognitive difficulties cited in Tapper’s book were public episodes that had been downplayed by Blue Team media elites.

Some experts hold that Edith was effectively the acting president during this period, usurping either her husband’s position or that of his vice president, Thomas Marshall.

Ronald Reagan, Hillary Clinton, Donald Trump, and Joe Biden are all examples of the parties violating the William Henry Harrison rule with their nomination. Until the election of Reagan in 1980, William Henry Harrison appropriately held the age record for a president.

Arguably, they tried to thin the field a little too quickly, and there were some unfair revisions of the eligibility criteria, but they could not (and did not) have the ability to determine which candidates did well in polls.

Democratic political elites were clearly not in favor of anti-establishment candidates, but were much slower to endorse candidates during the pre-primary season.

C.f. here.

Bernie Sanders remained a favorite among the alternative left media ecosystem.

All three had a track record of severe underperformance in statewide elections and bad head-to-head polling. Pete Buttigieg received more mixed coverage than the rest of this group, which can be attributed to the fact that while he was not straight and was very adept at navigating the social justice milieu, he was still a moderate from the midwest, and furthermore lacked the traditional qualifications to run for president.

Based on the differences in head to head polling, absolutely.

Within the Democratic Party of 2019, most of the leadership or figurehead talent pool with ambitions for higher office still consisted of straight white males. Additionally, racial minority politicians within the Democratic Party tended to come from safe states or districts.

E.g., Andrew Yang, Tulsi Gabbard, Marianne Williamson, arguably Tom Steyer.

Vice presidents usually don't win elections

Tomas McIntee — Sat, 31 May 2025 20:52:21 GMT

As I have discussed before, people tend to focus on the wrong thing when they consider how to choose vice presidents: In particular, many focus on electability (picking a running mate from a key state or region) or internal party politics (by nominating a rival from a different faction within the party). Vice presidents have three principal duties:

Serving as a replacement president in case of emergency.
Presiding over the Senate.1
Serving as a credible proxy for the president.

To date, there is no evidence that vice presidents have an electoral advantage in their home state or home region.2 On the flip side, when moving to the actual realities of performing the job of vice president, there have been multiple cases where picking a “balanced” ticket by selecting a vice president from another faction has backfired severely, most famously in the case of Andrew Johnson.3

Public domain image illustrating rather starkly why it’s important to pick a vice president on the merits.

Are there any electoral consequences at all?

I would assert vice presidential candidate is relevant if the choice is bad enough on the merits that it casts doubt on the wisdom of the presidential candidate in making that choice. Notably, I believe that choosing a running mate with substantially different politics and a thin resume (Sarah Palin) harmed John McCain’s bid for the president in 2008 because it made McCain look reckless. While the political science literature is divided on the subject, I am not alone in this regard.4

Public domain image of Sarah Palin in 2008.

John Edwards did not help John Kerry win votes in North Carolina in 2004, which Barack Obama subsequently won in 2008 without a North Carolinian running mate. Paul Ryan did not help Mitt Romney win votes in Wisconsin in 2012, which Donald Trump subsequently won in 2016 without a Wisconsonite running mate. While Hillary Clinton carried Virginia in 2016, it’s unlikely that Tim Kaine earned her any extra votes.5

There are a few cases where a running mate may have earned additional votes in their home state due to truly exceptional popularity or connections with a state-level political machine, but overall, vice presidential candidates rarely have a measurable positive electoral impact in their home state, much less their home region.6

In general, if there is an impact, it’s very small. One reasonable demonstration of this comes from the 1896 and 1900 elections, which both featured a contest between William McKinley and William Jennings Bryan. In 1896, William McKinley won 51.0% of the popular vote with Garret Hobart, a nationally obscure political insider who lacked the traditional qualifications for the presidency.7

Public domain image. Note that both candidates were usually known as “William” rather than “Bill,” as playfully put in this particular image.

In 1900, William McKinley’s running mate was a prominent and popular highly-qualified hero of the Spanish-American War, Teddy Roosevelt; McKinley earned 51.6% of the vote, a mere 0.6% higher than his previous performance.8 It is particularly worth noting that the ticket’s performance in Teddy Roosevelt’s home state of New York, where his impact should have been highest, declined from 57.5% to 53.1%; it’s not clear Teddy Roosevelt earned McKinley any extra votes at all.

Trump’s running mate, J.D. Vance

Trump selected Senator J.D. Vance of Ohio. As Vance was elected to the United States Senate only two years ago, his political resume is comparable to Sarah Palin’s in 2008. If Trump had died before being sworn into office, Vance would be sworn in with less relevant experience than at least forty-three out of forty-six presidents, arguably forty-five including Taft and Hoover.9 The only president to enter office with an indisputably thinner resume was Trump himself.

Unlike Sarah Palin, however, Vance was selected from the same wing of the Republican Party. Vance ran for office as a Trumpist. He has echoed Trump’s views on a wide range of topics. In addition, he has reportedly become good friends with two of Trump’s children. While Vance had a prior history of criticizing Trump before reinventing himself as a Trumpist, this is not unusual among Trumpists; very few qualified Republican politicians have a long history of loyalty to Trump.10

Vance is, therefore, a relatively credible proxy for Trump in spite of his thin political resume, and he seems as likely as any other qualified Republican to continue to execute Trump’s agenda in the event of Trump’s death or incapacity. Finally, the political cost to Trump of taking Vance out of the Senate was minimal: Vance was replaced immediately with a Republican appointee.11

Kamala Harris’s running mate, Tim Walz

Neither party is in a position to be able to afford to lose any seats in the US Senate if they want to advance their political agendas. The opportunity cost of nominating Tim Walz was near zero: He was a governor halfway through his second term in office, who, if elected, would be replaced by a Democratic lieutenant governor in a state that has not elected a Republican governor since 2006.

While not exceptionally popular, Walz was not exceptionally unpopular, either.12 More importantly, however, he was in alignment with Kamala Harris. There were very few voters who thought that, in the hypothetical death or incapacity of a President Kamala Harris, Vice President Tim Walz would steer the country in a different direction.13 This was not true of Josh Shapiro, Andy Beshear, or Roy Cooper, all of whom were generally identified with the moderate wing of the party.

Many other potential contenders whose names were floated were senators in vulnerable seats, such as Sherrod Brown, Raphael Warnock, and Mark Kelly. While not critical to a popular coalition, Josh Shapiro was a popular governor of a swing state in the middle of his first term, and thus recruiting him for the vice presidency would come with an opportunity cost as well.

When we consider all factors, Tim Walz was not a bad choice for Kamala Harris’s running mate: He was qualified for the role, he was a credible political proxy, and he was neither unpopular nor scandal-ridden. He did not gain votes for Kamala Harris, but it’s not clear that any vice presidential candidate could have helped Kamala Harris gain votes.14

Electoral calculus: Did it matter?

Vice presidents rarely, if ever, bring votes to the table. If they are exceptionally bad choices, they might cost votes, but neither J. D. Vance nor Tim Walz were exceptionally bad choices. Again, the main electoral criterion for vice presidential nominees is whether or not voters think they can do their job of serving as a proxy for their running mate, and both were reasonably qualified in that regard.

But political races are dynamic. Let me turn to the polls for a moment; there was a moment that the vice presidential candidates grabbed the limelight during their debate. By most accounts, Vance won; as there was no second Trump-Harris debate, this was the last major public debate. Did it move the needle? Well, if you look at the polls, the trajectory of the polling shifted a little bit after the debate. If the vice presidential debate was the only new factor in the last month of the campaign, we could generously conclude that Vance’s victory added up to half a point to Trump’s support.15

It’s possible - even plausible - that Vance’s superior debate performance tipped the needle a tiny bit.

Trump’s poll numbers were essentially flat when Biden was his opponent. When the perennially less popular Kamala Harris was nominated in Biden’s place, however, Trump saw a full one-point bump immediately in the polls, followed by a steady rise as more swing voters became familiar with Kamala Harris and consolidated behind Trump.

If, in other words, the only two events that mattered to change the trajectory of the race were the Democrats switching standard-bearers and the vice presidential debate, then the former event had five times the impact of the latter.

And that is an estimate that is very generous to the impact of the vice presidential candidates on the course of the presidential election. Detailed polling does not support the idea that many people were influenced by the vice presidential debate; on the other hand, polling highlighted the fact that voters were very aware of the more radical positions taken by Kamala Harris in her 2019 campaign, baggage that was unique to her.16

Vance is now vice president, making him an important figure in the Trump administration and likely to become president in the future. Walz is now considered a potential contender for the Democratic nomination in 2028, albeit a long-shot one. These consequences are far more significant than the dubiously small electoral impact of either selection.

That is not to say that VP choices didn’t impact the 2024 elections: The choice of Kamala Harris as VP by Joe Biden in 2020 was important in 2024, and helped sink Democrats’ chance of victory. As I discussed at the time, vice presidential candidates should be chosen based on credibility as a successor, not symbolic electoral concerns, because that’s the job of a vice president. Kamala Harris was not a credible proxy for Biden; she was a symbolic choice (selected for sex and skin color) from a rival wing of the party. Democrats have now paid the price for that mistake: If Biden had a more suitable vice president, he might have stepped aside earlier.

Democrats needed to have a high-quality emergency replacement for Biden; instead, they had a symbolically-chosen vice president from a different wing of the party, one who was less popular. The vice presidential choices during the 2024 campaign were ordinary and probably not at all impactful in terms of electoral calculus; the most important facet of the choice is that Vance is now Trump’s designated successor.

Since the time of LBJ, this has been treated as a largely ceremonial duty with minimal power beyond an occasional tiebreaking vote.

The number of presidential elections is small, so this doesn’t mean that there isn’t one.

The basic issue is that it’s not actually unusual for presidents to die or become incapacitated, and political assassins have a much stronger motive if the vice president differs politically from the president.

C.f. this paper in favor of my position and this paper arguing against it.

Obama won Virginia by 4% in 2012, Clinton by 5% in 2016, and Biden by a whopping 10% in 2020.

For possible exceptions, I would highlight LBJ’s deep connections to the political machine in Texas in 1960. Another potential recent example would be Obama’s performance in Delaware in 2008 and 2012 with Biden on the ticket.

He served in the New Jersey state legislature for about twenty years. While doing so, he devoted most of his time to a lucrative law practice, treating politics as a part-time hobby rather than a career.

Teddy Roosevelt had served first as Assistant Secretary of the Navy and then as Governor of New York; at the time, New York was both the largest state and a crucial political battleground state.

Herbert Hoover (elected 1928) was Secretary of Commerce for eight years; Taft (elected 1908) was the unelected military governor of the Philippines and Secretary of War.

This included essentially the entire field of 2016 Republican politicians prior to Trump’s nomination, all the members of the first Trump administration that ended their relationship with Trump on contentious terms, and then a significant number of Republicans who favored nominating someone else in 2024.

Ohio has a Republican governor.

As Kamala Harris was.

If you asked mainstream Democrats, they would identify both as progressives. If you asked hard leftists or Republicans, they would identify both as hollow suits, faces for The Blob of Democratic machinery. Walz’s agenda as governor of Minnesota was largely in alignment with the agenda of the Biden-Harris administration. Kamala Harris herself was a bit of a political chameleon over her political career, ranging from more moderate to an outspoken member of the Horseshoe Caucus.

Once again, vice presidential candidates rarely have any positive effect on the ticket, even in their home state.

This is not plausible. It is normal to see an accelerated shift (“October surprise”) as less-engaged voters tune in late, and Trump did very well among low-propensity voters.

It is very common for a significant share of voters to decide late in the campaign, and there certainly were other things that happened during October of 2024. Some of Kamala Harris’s specific baggage as a candidate was highlighted as “misinformation” about the Democratic platform, such as by Blueprint; however, almost all common “misperceptions” about Kamala Harris’s platform reflected previous statements made by Kamala Harris. A few of those (such as border control) also reflected concerns arising from the de facto practices of the Biden administration.

Meritocracy, mediocrity, & midwits

Tomas McIntee — Sat, 17 May 2025 18:46:23 GMT

Democrats lost narrowly but decisively in 2024 to the party of Donald Trump, a loss fueled by “double haters”: anti-Trump voters who decided that the Democratic Party was the greater threat. I’ve been watching carefully to see if the Democratic Party will change course; I don’t think it will happen.

The deeper problem is that the Democrats have, as a party, rejected meritocracy, both at a level of policy and at a level of internal governance. This has put them at odds with the median voter and also left them without effective leadership.

I’ll explain this image later.

It’s accurate, if crude, to say that the Democratic Party, as an organization, has become a party of so-called “midwits,” a pattern I feel has been firmly confirmed in the recent DNC elections. Past successes of the Democratic Party have been based on highly intelligent candidates skilled at connecting with blue collar workers.1

The scope of the problem

In the last 33 years of Gallup polling, a plurality of the population has leaned towards Republicans for only three years: 2022, 2023, and 2024. Quinnipiac has just registered the highest disapproval rating for the Democratic Party in the 17 years that they have been polling the question. The post-Obama leadership of the party has low approval ratings.2

Only Donald Trump could make Hillary Clinton look popular in 2016.

From 2015-2024, Trump himself was deeply unpopular outside of a brief honeymoon period at the very start of his first term.3 The Republican Party was also generally unpopular in the entire post-Reagan era, which has helped mask the degree to which the Democratic Party has lost its way.4 The roots of Democratic divergence from popular positioning date to the pre-Trump era, but didn’t necessarily become visible immediately.5

There are two major processes that have taken place in parallel. First, conservative and moderate Democrats have gradually left the party. Second, Democrats have rapidly adopted a collection of postmodern beliefs and values.6 One of the underrated accompanying effects is that Democratic activists and elites now tend to think using methods akin to literary analysis, focusing on symbolism and hidden authorial intentions.7

Meritocracy & mediocrity

Meritocratic processes have been perennially popular in America except for certain key interest groups, most of which are aligned with the Democratic Party.8 Affirmative action policy has been pushed forward by Democrats in response to the narrow interests of those key interest groups, traditionally thriving on obscurity and framed carefully in liberal terms as the prevention of accidental discrimination.9

This has always been a losing position in the court of public opinion, and also in the court of law. In spite of the legal and political jeopardy, postmodern progressives embraced increasingly overt discrimination under the label of DEI, leading to a number of successful anti-discrimination lawsuits against progressive institutions and the erasure of the narrow legal exceptions that had allowed affirmative action to continue.10

A landmark SCOTUS case. Remixed from source with fair use of identifying logos.

A number of Democratic elites sincerely believe that there is no tension between hiring for competence and symbolic inclusion, which has led them to prioritize symbolic inclusion for everything from hiring staff to selecting political candidates. Acting in accordance with these beliefs has systematically reduced the competence of Democratic-aligned organizations, volunteers, staff, and candidates.11

A party of midwits?

Famous midwit meme (left) versus actual typical position of so-called “midwits” on the IQ curve (right). Most “midwits” are actually smart people around the 80th-95th percentile. It’s not a very useful term.

In most domains (and politics is no exception), those who succeed at a high level purely on merit are usually highly talented; however, people who are merely moderately talented can succeed if they have some form of privilege.12 Increasingly, Democratic leadership is full of people who are moderately talented rather than highly talented — or, in Twitter parlance, midwits — because symbolism has been deemed more important than talent.

A recent prominent example is Kamala Harris, the 2024 Democratic nominee for president. In order to quell complaints about the nomination of a white man for president, Joe Biden promised to pick a woman for vice president, then a black woman; out of the extremely limited pool of options left after making those public promises, he chose Kamala Harris. When Joe Biden gave up the Democratic nomination, Kamala Harris was the only plausible backup choice.

Objectively speaking, Kamala Harris was not without talent, and likely had an above-average IQ.13 However, her main strengths were in “inside baseball” politics, not winning elections, and her electoral talents were not exceptional.14 She became the party’s figurehead because of her symbolic qualifications rather than her abilities.

On the other end of the spectrum (moving from figureheads to operations), I will discuss two prominent new officers of the DNC, elected in 2025: David Hogg and Malcolm Kenyatta.15 While both could be more talented at electoral politics than the average non-politician, both have risen to prominence for symbolic reasons rather than political talent or intellectual brilliance.

Malcolm Kenyatta’s statewide electoral track record consists of two failures: In 2022, he lost a Democratic primary by a large margin to a less well-connected candidate.16 In 2024, he made it to the general election butunderperformed Kamala Harris by a significant margin, posting one of the worst performances of any battleground state Democrat running in 2024.17 His most notable achievement is that he was the first openly gay black man in the Pennsylvania state legislature.

David Hogg is a modern digital-age political activist. While bright and talented enough to reach a notable social media following,18 his rise to prominence and rise within the Democratic Party has been largely due to his symbolic status as a prominent school shooting survivor. I decided to feature him in this article as an example of a midwit back in February of 2025; since that time, he has:

He’s bright. He’s right about some things, in fact; Democrats have a gerontocracy problem and are being weighed down by unpopular incumbents.19 Democrats have a very visible sexism problem that is driving away young men.20 Democrats are the party of scolding women and arcane rules designed to penalize and punish young men, and of selective enforcement of rules on, in particular, young, white, heterosexual men.

Clips from the DNC election circulated widely on Twitter, particularly the ones of DNC officers tripping over trying to read and explain the gender quota rules.

And now, the DNC has decided, retroactively, that their race and gender balance quotas for officer positions were violated in electing David Hogg, in spite of the fact that they had read aloud and explained those rules on live television during the election process. So, to recap:

Democrats made a show of adhering diligently to unpopular norms surrounding affirmative action in the DNC elections, because symbolism is more important than competence.
They then chose (unwisely) a vice chair on the basis of his symbolic importance.
He then correctly but unwisely started an internal fight over incumbents, and also pointed out that they had a number of problems, including driving away men… especially young and heterosexual ones, which has a lot to do with capricious and arbitrary enforcement of complex rules designed to tip the balance against white men.
The DNC then capriciously and arbitrarily voided his election retroactively, based on rules that were originally designed to tip the balance against white men.

Every part of this story revolves around Democrats turning away from meritocracy, resulting in “midwit” grade mistakes: The complicated gender balance rules were clever but unwise. Electing David Hogg was clever but unwise. David Hogg’s off-the-rails behavior was all clever but unwise. Choosing to get rid of David Hogg by voiding his election on the basis of his sex was, again, clever but unwise.

All the way up

Out of the Democrats’ last six nominees for president, two, Bill Clinton and Barack Obama, won the nomination the hard way. They had gained national prominence in the first place thanks to notably superior electoral performances, and displaced better-connected candidates who had been considered early frontrunners.

The other four were “inside track” candidates who rode into the nomination on the coattails of others and did not perform as well: Al Gore (as Bill Clinton’s vice president), Hillary Clinton (as Bill Clinton’s wife), Joe Biden (as Barack Obama’s vice president), and Kamala Harris (as Joe Biden’s vice president).21 As meritocratic processes weaken, the social inside track of nepotism tends to strengthen.22

Americans love blind justice and liberty. Meritocracy is as American as apple pie; we didn’t invent it, but it’s a long-time favorite.

More meritocratic selection processes for nominees mean stronger candidates. More meritocratic selection processes for back-office types (ranging from campaign staff to the DNC officers) means more competent support for candidates. It’s hard to overestimate the total cumulative impact on Democrats of fully embracing symbolism over merit from top to bottom has had on the competence of the Democratic Party.

If the Democratic Party is to become both popular and capable in the future, it must embrace meritocracy, both as a matter of public principle and as a matter of private practice. I do not expect the Democrats to change course quickly; at present, the anti-meritocratic factions have dominated elite circles for at least a decade.

The “Obama-Trump” voters were very much working class. Bill Clinton was no less striking of an example of extraordinary intelligence combined with the ability to connect with the working class. Democrats do still have significant support among African-Americans in lower socioeconomic strata, although even that has declined.

Obama still posts positive net favorable numbers. Hakeem Jeffries has net positive approval rating in the latest Gallup poll… but that’s still with only 35% positive ratings, suggesting his net positive approval is only due to the fact that he’s relatively obscure and has yet to work with a House majority.

Trump’s first-term approval ratings were positive until about two weeks into his term in office. This time around, he was in net negative territory until he took office and started signing executive orders. It remains to be seen whether Trump’s current net positive approval ratings will prove to be a temporary honeymoon period or not; there are very significant differences.

See this Gallup poll series for a 1992-2021 trend in party favorability and again this Gallup poll series for party identification showing that Democrats enjoyed an edge in identification/partisan leaning from 1992-2021, with Republicans overtaking Democrats in 2022-2024 after last being up in 1991. Note also that Democrats won the national popular vote in seven out of eight presidential elections from 1992-2020, losing mainly in midterm and off-year elections with lower turnout.

Obama was personally very popular in 2012, when the Democratic platform began a dramatic shift influenced by the popularization of various postmodern positions through social media.

This cluster may be referred to as “applied postmodernism,” “wokism,” “progressive memeplex,” or various other terms. It includes distinct identifiable components that have been in the spotlight at different times, but that overlap significantly. For example, the Critical Race Theory movement was in the spotlight in 2018-2020 thanks to works like White Fragility, The 1619 Project, and How to Be an Antiracist. When Hamas started the war in Gaza in 2023, postcolonialism moved into the spotlight, but many of the same organizations and individuals were involved in activism, e.g., BLM chapters posting paraglider images in support of Hamas’s paraglider attacks of October 7th. While postmodern activist movements tend to reject clear labels and definitions, the existence of the cluster is clear.

The reasons for this are not entirely clear. However, postmodernism’s roots draw heavily on techniques used for the critique of art and literature, and postmodern thought tends to be hostile to modernist projects, the Enlightenment, rationalism, analytic philosophy, science, et cetera. Additionally, education in the humanities and related fields has both become more heavily aligned with partisanship and more common among elites. You can see this particularly in the sharp increases of the use of the phrase “lost the plot” to refer to real-life events, or jokes about “the showrunners” that frame the real world as a television show.

Notably, African-American interest groups (affirmative action) and educator interest groups (opposition to standardized testing).

Polling on the topic of “affirmative action” is sometimes positive, polling describing how affirmative action functions inevitably shows steep opposition. Notably, Californian voters have repeatedly rejected affirmative action at the ballot box even while voting for Democrats in large numbers.

Most notably SFFA v. Harvard,

There has always been a lot of nepotism in politics, which compounds the problem; affirmative action hiring doesn’t displace nepotistic hires typically, which means that it impacts merit-based hires.

Traditionally, privilege in political organizations meant social connections, particularly but not only familial connections. For example, in 2000, both major party candidates had fathers with prior political careers.

While she failed the bar exam on her first attempt, the pass rate on the California bar when she took it was roughly 45%. Her trajectory of a failed attempt followed by later success was very ordinary among contemporary California lawyers, neither a sign of unusually good nor unusually bad capabilities.

This is not unusual for “safe state” Democrats, where winning endorsements can be a quick route to near-inevitable victory. No Democrat has performed worse against a Republican statewide in California from 2010 to the present than Kamala Harris in 2010. She managed performance near the middle of the pack in 2014 (subpar if one adjusts for incumbency), faced a fellow Democrat in 2016, and had a presidential campaign disintegrate in 2019 with rock-bottom polling against Trump.

The other DNC members are traditional “back office” operators who do not have the same kind of existing public profile; these two, however, are aspiring “front men,” and the DNC elections of 2025 were a highly public affair broadcast live on cable television.

Notably, Kenyatta received more endorsements than Fetterman, but placed a distant third, with Fetterman winning every single county in spite of a lack of insider support.

The worst performance belonged to Pennsylvania’s Democratic nominee for Treasurer, who ran on an explicitly anti-Israel platform and also was explicitly disowned by Pennsylvania’s popular Democratic governor. While Kenyatta did at least manage to win the 2024 primary for Auditor General, his only opponent was a nearly-unknown county controller who had no notable endorsements and previously lost two races for the state senate.

While often derided as a “dumb kid” by critics online, Hogg’s SAT score was in the 83rd percentile and he clearly has a knack for social media. He’s not a genius, but he’s brighter than average. At present, he has roughly 1 million Twitter followers and 200,000 Instagram followers, roughly comparable to John Fetterman but significantly behind Alexandria Ocasio-Cortez.

Not the ones on his target list, necessarily. Nancy Pelosi, for example, is deeply unpopular. Some of the Democrats’ dead weight can be found among their younger members, though, like Rashida Tlaib and Ilhan Omar, both of whom have caused serious damage to the party with their antics.

Notably, in the last election, including a substantial number of minority men, though Democrats also have a problem with overt anti-white racism.

Joe Biden and Al Gore both did make serious political bids of their own, and earned the highest share of the

I talked about this in my last article.

Why affirmative action failed

Tomas McIntee — Fri, 27 Dec 2024 17:33:48 GMT

It’s been over a year since the Supreme Court banned racial affirmative action in Students for Fair Admission v. Harvard (2023). The vast majority of elite universities within the United States pursued aggressive affirmative action policies for a period of over half a century, a practice affirmed with limits in Regents v. Bakke (1978).

While one of the last exceptions to anti-discrimination laws was closed in the SFFA v. Harvard case, it is difficult to prove and prosecute discrimination, and those who believe that affirmative action is productive will continue to do so for some time in defiance of the ruling.1 Neither opponents nor proponents believe that affirmative action will end short of legal force, even as it becomes clear that the ideological frameworks used to justify a system of perpetual affirmative action create discord and division.2

Yet when affirmative action began, it was nearly universally described as a temporary measure by proponents - a one-time push to overcome decades of discrimination. After almost sixty years of affirmative action policy, it’s time to admit that affirmative action failed to accomplish its mission and ask why it failed.

Enough time has passed to declare failure

Supreme Court. Distance between left and right exaggerated for effect in discussion of this 6-3 decision.

The Supreme Court that decided Bakke was one that was not personally familiar with affirmative action. The Supreme Court that decided SFFA v. Harvard was a court that could not remember life before affirmative action. The oldest justice was Clarence Thomas, an African-American justice who voted against affirmative action. When Clarence Thomas was admitted to Yale Law School in 1971, the Ivy League school was operating under an aggressive affirmative action program with a target quota of 10% of the law school’s entering class.

Sex-based discrimination within higher education became illegal in 1972; by 1990, women earned more bachelor’s degrees than men.

Enormous changes have taken place both in the workforce and within higher education, including the reversal of the gender education gap and the elimination of gender discrimination in pay.3 Similar changes have not emerged in the domain of race. While individual African-Americans have broken through to the highest levels of the elite,4 population-level gaps in income, wealth, and educational attainment remain very substantial, with few signs of overall progress in the last two generations.5

In fact, progress on racial income gaps, wealth gaps, and educational performance gaps between white and black Americans has been slower than would have been expected simply by the removal of discriminatory barriers. Since racial affirmative action was justified as a method of remedying recent historic discrimination against African-Americans under Jim Crow law, the failure to close the gaps between white and black Americans is a failure of affirmative action to accomplish its mission.6

Meritocracy works

As with many psychological and social phenomena, at most half the variation in socioeconomic status in the United States appears to be inherited from parents.7 A simple model of the impact of the removal of discrimination would predict an exponential decay in the impact of historical discrimination on each successive generation, with around 90% of the impact being erased by the third generation. This may seem slow, but we are now looking at the third generation of affirmative action recipients and have not seen anywhere near this kind of progress.

Historically, this crudely corresponds to what we see in immigrant communities.8 First-generation immigrants from non-English-speaknig countries face real barriers to integration; second-generation immigrants are fluent in both the English language and American culture; and the large majority of third-generation immigrants are difficult to distinguish from “old stock” Americans outside of holidays and family reunions. Very little scholarly work even attempts to use the concept of “third-generation immigrant.”

Even during periods where de jure discrimination was openly practiced, rigorously meritocratic processes coincided with significant socioeconomic progress by minority communities. The most significant closure of the gap between black and white wealth and income (circa 1870-1910) coincided with the presence of an integrated and strictly merit-based federal civil service system (1883-1913).9

Similarly, the first rise in standardized college admissions exams circa 1901-1921 offered Jewish-Americans chance to break into traditionally WASP-dominated elite social circles via Ivy League education.10 Even a small amount of strict meritocracy can be very effective at remedying the impact of of discrimination, particularly against a minority group. Why? To answer this, I’ll turn to game theory.

Discrimination is an anti-coordination game

One of the many ways in which companies compete with one another in a free market is in hiring. This extends naturally to college admissions, awarding of grants, and other settings in which affirmative action has been utilized.

For the sake of simplicity in modeling, I’ll talk about an example with a single finite hiring pool of candidates that vary in a single dimension of quality, and two companies. The first company is large, well-resourced, and acts first.11 Candidates come from two groups with identical distributions of ability.

Let’s say that the first company engages in discrimination, but is also able to imperfectly observe ability. They hire what they believe to be the best members of Group A, while ignoring candidates from Group B regardless of how well they interview.12

It is now Company 2’s turn to hire. What does the hiring pool look like? Well, it includes a lot of high-quality members of Group B that were discriminated against by Company 1, and the members of Group A still in the hiring pool tend to be lower-quality, since this includes the candidates rejected by Company 1 for a poor interview score.

Consider three strategies for Company 2:

Company 2 also discriminates in favor of Group A.
Company 2 hires on merit (default strategy).
Company 2 discriminates in favor of Group B.

For the three extremal strategies, the distribution of hired candidate abilities looks like this:

Hired candidate distribution of ability if Company 2 discriminates in favor of Group A (violet), in favor of Group B (blue), or does not discriminate at all (gray).

If Company 2 discriminates in favor of Group A just like Company 1, the quality of their hires suffers substantially.

If Company 2 discriminates in favor of Group B, the resultant quality of their hires is almost unchanged.13

Affirmative action can have a small positive effect on hiring pool quality in theory, but only if discrimination in the other direction is extremely pervasive. Even then, the benefits from abandoning strictly merit-based hiring tend to be extremely small and fragile, evaporating as soon as other companies adopt merit-based hiring and reversing if other companies adopt similar affirmative action strategies.

Additionally, what’s very clear is that the more companies engage in affirmative action in favor of Group A, the better the results for companies that persist in discriminating against Group A. Not only does discrimination breed discrimination on a tit-for-tat identitarian basis, discrimination enables discrimination as a game theoretic exercise—because discrimination is an anti-coordination game.

The leaky pipeline & the backlash

The core problem that affirmative action sought to address was that even after applying relentlessly blind meritocratic techniques to processes like hiring or college admissions, qualified candidates were still disproportionately unlikely to be African-American. While desegregation brought a decline in gaps in educational achievement, there are signs that the racial gap in K-12 educational achievement have increased since 1988.14

NAEP long-term trends in the black-white gap for 13 year olds in math and reading. Note transition in testing format makes it hard to compare data before and after 2004 directly, but an increasing gap is visible both in the 1988-1999 period and the 2008-2023 period.

Affirmative action in hiring and college admissions were a way of achieving temporary results (greater visible representation on college campuses and in prestigious jobs) without fixing the underlying problem (achievement gaps that start in K-12 education, leading to a gap in qualified candidates). This backfires in three distinct ways:

The double standard reinforces a stereotype of an achievement gap.
Underqualified students fail out of challenging programs.
Double standards generate resentment.

A selective college that admits African-American students on the basis of lower standards of academic performance creates a class environment in which their students and professors experience a reality in which African-American students are less prepared and capable.

It is an environment in which negative stereotypes are accurate. It is an environment in which professors must either fail out African-American students at higher rates, grade them on a different curve, or lower the standards of instruction until few if any students fail at all.

The nepotism problem

We can’t talk about college admissions or the job market without talking about nepotism. Ivy League schools and other elite universities are known to give special consideration to legacy students and children of donors.15 Legacy admission was widely formally adopted by the Ivies as part of an effort to limit the influx of high-performing Jewish students in the 1920s.16 The problem was that Jewish students persistently scored better on entrance examinations.

In 2019, a major scandal broke: Many of the rich and privileged were cheating to get their kids into elite schools in a $25 million entrance exam cheating ring. Wealthy parents cannot simply buy a good SAT score.17 In 2020, a number of colleges and universities declared that they were dispensing with the SAT, a measure inspired by COVID-19. This was loudly celebrated by affirmative action advocates… and likely much more quietly by privileged parents whose children did poorly on the SAT. Proctored examinations are not perfect, but remain one of the best objective measures of talent.

US Navy sailors preparing for the ASVAB. Why? Because the military wants to ensure a level playing field and meaningful test results.

While nepotism and affirmative action are sometimes discussed in terms of balancing factors in a calculus of racial politics,18 they have a synergetic relationship: Nepotism thrives in systems that are opaque, relying on subjective judgement of hard-to-quantify factors. Since explicit quotas and score-based systems providing a transparent advantage by race are strictly illegal, affirmative action requires non-transparent systems based on subjective discretion.19

Additionally, many negative impacts of nepotism and affirmative action are also synergistic. Both nepotism and affirmative action put downward pressure on quality standards, for example, leading to grade inflation. Both reduce social mobility and efficiency by limiting broad-based meritocratic access to elite institutions.

Finally, many of the individual beneficiaries of affirmative action come from economically privileged backgrounds, who know to highlight trace or non-existent minority heritage or who come from a global elite. They are identified with oppressed Americans who lived under Jim Crow laws symbolically rather than materially. For example, the now-infamous Claudine Gay came from a wealthy family that owns Haiti’s largest concrete plant and went to an elite boarding school; neither Gay nor even her ancestors were affected by Jim Crow.20

Recap and review

After three generations, there is no reason to expect race-based affirmative action to create any meaningful closure of any educational or economic gaps. Affirmative action as practiced widely and openly by both private and public universities for more than half a century is now thoroughly illegal, as is affirmative action in the corporate world. The Supreme Court had always viewed this as a temporary remedy at best, and affirmative action has decisively and clearly failed to meaningfully close population-level gaps between black and white Americans.

Even in the presence of widespread discrimination against a group, strictly merit-based selection is usually better for the institutions using it, and institutions adopting fashionable types of affirmative action can expect to reduce the quality of their selections as a result.

It is long past time that institutions chose to judge people by the content of their character rather than the color of their skin. (To preempt the inevitable arguments: While Martin Luther King Jr. favored what he called affirmative action in Why We Can’t Wait, he described a system that would benefit poor whites, i.e., economic rather than racial.)

In addition to the failure of affirmative action to remedy the population-level gaps between black and white Americans, affirmative action has numerous second-order negative effects. Affirmative action supports and creates stereotypes of incompetence by people labeled as “affirmative action hires.” It also obscures the presence of urgent root problems that need addressing earlier in the pipeline, particularly in the domain of K-12 education.

Finally, the obscurity and reduction in emphasis on objective merit demanded by affirmative action processes also benefit well-connected and privileged candidates, as nepotism thrives in the darkness.

Notably, Yale and Princeton have released demographic data showing that the share of Asian-Americans admitted declined. Since the type of affirmative action practiced by Harvard and other Ivy League schools was found in particular to discriminate against Asian-American applicants, this suggests that these schools continue to discriminate against Asian-Americans.

It is also worth noting that while universities had an exception, this exception did not extend to hiring, but affirmative action (that is to say, discrimination in favor of groups seen as underrepresented) has remained commonplace in some sectors of the job market for decades.

See, for supporting evidence, the New York Times’s recent lengthy piece on DEI at the University of Michigan. The position that race-based affirmative action would be required on an indefinite basis would have been completely untenable by the ideological standards employed by the majorities in the Warren and Burger courts in the early key decisions.

Remaining gender gaps in income are related to career field, hours worked, educational attainment, and career duration; women work fewer hours, retire earlier, and tend choose fields with benefits and security rather than higher nominal salaries. Teachers, for example, are very poorly compensated in terms of nominal annual salary, but have high total compensation per hour. In some subgroupings, women earn more than men in raw compensation, explicable mainly by higher levels of educational attainment.

In the executive branch, the highest color barrier was broken in 2008 with the election of Barack Obama; in the legislative branch, there are about sixty African-American members of Congress, roughly proportionate to the African-American share of the population. In the private sector, there are ten African-American billionaires and eight African-American CEOs of Fortune 500 companies.

For example, the gap in home ownership rates by race increased from 1980 to 1990 and appears to have changed very little in the past 100 years; restricting to male-headed households in order to eliminate the effects of a separate pattern of a rising marriage gap shows the gap reached a minimum among comparable types of household around 1980, roughly one generation after the removal of formal discriminatory barriers began and two generations before the present time.

See in particular Bakke (1978) and the discussion therein about “remedial” preferences.

See, for example:

https://doi.org/10.1086/378526
https://doi.org/10.2307/2553676
https://doi.org/10.1007/s10888-019-09413-x

The major exception to this pattern of large-scale integration within three generations include religious groups with substantial lifestyle differences and racial minorities subjected to formally discriminatory policy, such as the Amish, a wave of early Chinese immigration that was formally curbed by the Chinese Exclusion Act in 1882, and a wave of early Japanese immigration that was informally curbed in 1907 and formally ended in 1924. Asian-Americans largely ceased to be an exception to this pattern following the removal of de jure racial discrimination.

Note that the period between desegregation of federal employees, begun in 1948, and the elimination of merit-based civil service examinations following a 1981 court case, marked another period of closing of this socioeconomic gap.

Harvard began actively working to reduce Jewish admissions in 1922.

This is equivalent to having a large number of companies that hire first.

Simulation details: Interviewed Quality as the sum of three normally-distributed Ability (mean = 50, sd = 15) + Noise (mean = 0, sd = 15). Pool of n=20000 cloned candidates whose only difference is group membership. Initial hire of 5000 candidates by Company 1, follow-up hire of 2000 candidates by Company 2.

Whether there is any advantage at all to discriminating in favor of Group B depends on parameters.

NAEP is one of the best long-term direct measures of the K-12 system, but similar patterns can be seen in SAT scores and other less direct measures. Note that 1988 would be 12 years after Runyon v. McCrary and 34 years after Brown v. Board; minimum racial gap corresponds to parents who had access to integrated public schools with children who entered school after Runyon v. McCrary banned private segregation academies.

As discussed in the SFFA v. Harvard / UNC cases, “legacy” admissions processes are even used in some public universities.

This was also accompanied by a reduced emphasis on admissions tests, geographic quota systems, et cetera—familiar features that have returned in the modern affirmative action era.

I could say a lot about the SAT—it’s a test I first took in 5th grade, so I had an unusual personal experience with it. However, the basic relevant point here is that the best preparation for the SAT is a good education in the fundamentals, with SAT-specific prep techniques having a marginal effect on top of major effects from skills in reading, reasoning, and mathematics. The best efforts of rich parents to improve their kid’s 1100 score are unlikely to push it to a 1300… without paying a ringer to illegally take the test in their place, at least.

Legacy admits in particular tend to be white, and a raft of “tit-for-tat” articles in progressive media called for the banning of legacy admissions following the SFFA v. Harvard case.

See Regents v. Bakke (1978), Gratz v. Bollinger (2003).

Preferential hiring for jobs via affirmative action was already illegal when Bakke (1978) was decided in favor of affirmative action for admissions purposes, but it is worth noting that affirmative action in favor of Claudine Gay would not have advanced any meaningfully remedial purpose.

Did Harris lose because of Gaza?

Tomas McIntee — Sun, 01 Dec 2024 20:06:41 GMT

Among the online left, it has slowly become fashionable to say that Kamala Harris lost because she did not support an arms embargo against Israel and did not make peace with Rashida Tlaib’s anti-Biden movement. This is not only wrong, it gets the impact of the war in Gaza on the election completely backward: If anything, Kamala Harris lost votes because she was seen as too friendly to the pro-Palestinian factions.

Now, granted, most American voters do not think that foreign policy is very important, but:

Most Americans side with Israel over Palestinians.
Most Americans strongly disapproved of the October 7th attacks.
The pro-Palestinian protests were harmful to the Democratic brand.

Left: Senator Kamala Harris meeting with Palestinians in the West Bank. Right: Vice President Kamala Harris meeting with Benjamin Netanyahu.

Kamala Harris ran as a moderate on the war in Gaza. If she had bluntly told the pro-Palestinian protest movement to go pound sand and taken an unequivocally strong pro-Israel stand, it probably would have won her more votes. Both the scale and the popularity of the pro-Palestinian protest movement have been frequently widely overestimated.

Israel is broadly popular in the US

The most recent Gallup poll testing favorability of Israel showed that 58% of Americans view Israel favorably and 36% view Israel unfavorably. With a net favorable rating of +22%, Israel is viewed more favorably than any living American politician.1 Pro-Israel views dominate the center of the political spectrum where most swing voters live, with anti-Israel views mostly relegated to the left and right fringes.2

As recent events have highlighted his strong pro-Israel views, John Fetterman (img src) has gone from being one of the least popular senators to positive net favorability, with a Quinnipiac poll showing his views on Israel to be driving a positive shift in public opinion. He wasn’t up for re-election in 2024, however.

As an elections guy, I will now step through ten key presidential battlegrounds to talk about how this played out in actual voting behavior. Voters have generally rewarded staunchly pro-Israel candidates in statewide elections in battleground states, while anti-Israel political positions and antisemitic rhetoric tend to lead to poor electoral performance.

Minnesota

During the Democratic primaries, Minnesota was Biden’s weakest state, with strong showings by his lone conventional Democratic opponent in his home district (MN-03) and high rates of “uncommitted” votes among Muslim voters in Minneapolis.3 Minneapolis is home to Ilhan Omar, a Somalian immigrant, one of three current Muslim members of Congress, and a vocal opponent of Israel.

Left: Amy Klobuchar among a select delegation of senators visiting Israel. Right: Ilhan Omar and “the Squad.”

Among the states that Kamala Harris won, this was the second-closest, with a margin of 4.24%. Was this low margin due to her failure to placate the anti-Israel vote? Minnesota is also home to Amy Klobuchar, who voted in favor of banning the BDS movement and has been repeatedly endorsed by AIPAC. On the same ballot as Kamala Harris, Amy Klobuchar won by a decisive landslide margin of 15.70%.4 Whatever support Klobuchar may have lost among anti-Israel voters, she more than made up for it with a broader appeal to moderates.

New Hampshire & Maine

New Hampshire was Kamala Harris’s closest-won state with a 2.78% popular margin. The New Hampshire primary was the most transparent electoral test of the Gaza protest movement, with a widely-publicized movement to write in “Ceasefire.” A grand total of 1,546 “Ceasefire” votes were recorded, including 34 in the Republican primary; about 1.2% of the Democratic primary electorate and about 0.3% of the overall electorate. After this weak showing, the Gaza protest movement rebranded itself as “Uncommitted.”

“Ceasefire” placed seventh in the 2024 NH Democratic primary, between Derek Nadeau (1616 votes) and Vermin Supreme (912 votes, pictured above)

While 4.3% of the Democratic primary electorate cast no preference (uncommitted) in 2024, this was less than the 5% of the Democratic primary electorate that cast uncommitted ballots in 2012, suggesting that nationally, the overwhelming majority of uncommitted ballots were cast for other reasons.5 It is particularly worth noting that the share of ballots that were uncommitted in 2012 reached 10.7% in Michigan and 20.2% in North Carolina.

Meanwhile, next door in Maine, the Harris campaign lost one electoral vote in Maine’s second district with a 9.0% popular vote loss while a staunchly pro-Israel Democratic candidate endorsed by AIPAC, Jared Golden, won re-election. This was Harris’s 8th closest contest, just behind battleground state Arizona.

Wisconsin & Michigan

Kamala Harris’s narrowest two losses, by margins of 0.87% and 1.41%, were Wisconsin and Michigan. These were the two states where the uncommitted vote6 during the primary was higher than the raw margin of loss: 48,000 versus 30,000 in Wisconsin, or 100,000 versus 80,000 in Michigan.

However, as mentioned above, not all uncommitted votes were cast in support of the “Uncommitted” movement. A better proxy for the Gaza protest vote in the general election was the vote for anti-Israel third party candidates, which added up to about 50,000 in Michigan and 15,000 in Wisconsin.7 It is worth noting that the Arab-American community in Michigan has been trending Republican due to social conservatism and showed stronger support for Trump this year than in 2020, but it is difficult to parse votes for Trump as meaningfully prioritizing Palestinian interests.

Left: Rashida Tlaib plugs a book by Linda Sarsour (condemned by Biden) while wearing a pro-Palestinian T-shirt with keffiyah & a map erasing Israel. Right: Elise Slotkin at the Ukraine Lend Lease Act signing, on good terms with the Biden administration.

While it is plausible that some unwisely-cast Gaza protest votes were topline undervotes or were cast for Trump instead of a third party proxy, it is worth noting that a Jewish Democrat with a more vocally pro-Israel record was on the ballot statewide, running for Michigan’s Senate seat.

Elise Slotkin won by 0.38% of the vote, running 1.79% ahead of Kamala Harris. Slotkin, a fellow member of Michigan’s Congressional delegation, has been highly critical of Tlaib. She performed poorly in Dearborn compared to Green Party candidate Douglas Marsh, whose #1 signature issue was a pro-ceasefire campaign. Marsh earned 54,000 votes across the state, outperforming Jill Stein; in other words, Slotkin faced a slightly larger Gaza protest vote as Kamala Harris while winning the election. Slotkin more than made up for this in the rest of the state, which is not aligned with its Arab-American minority on this issue.

Pennsylvania

Pennsylvania was the pivotal state of the 2024 election. Kamala Harris lost Pennsylvania by a margin of 1.75%. On the same ballot, Senator Bob Casey had earned the endorsement of AIPAC and had been critical of the Biden administration for being insufficiently pro-Israel; he ran about 1.5% ahead of Kamala Harris. It’s also worth noting that in the 2022 midterms, Pennsylvania voters elected Josh Shapiro and John Fetterman, two staunchly pro-Israel Democrats.8

Left: Anti-Israel candidate Erin McClellan’s support in the 2024 election. Right: Pro-Israel candidate Bob Casey’s support in the 2024 election.

Pennsylvania also saw a more direct test of the hypothetical question of “What if a Democrat embraced the Gaza protest voters?” The issue of state investment in Israeli bonds was the most heated political issue in the treasurer’s race, with Democrat Erin McClellan campaigning on divesting from Israel. On the same ballot as Bob Casey and Kamala Harris, she lost by a margin of 6.3%, slightly worse than Malcolm Kenyatta.9

Georgia & North Carolina

The Gaza protest movement was more clearly a non-factor in southeastern swing states. Kamala Harris lost Georgia and North Carolina by 2.20% and 3.10% of the total vote, respectively, with Trump winning an outright majority of the vote. Only about 6,000 of Georgia’s Democratic primary voters chose to cast a blank ballot; while 12.7% of North Carolina’s primary electorate cast “No preference” ballots, this was lower than the 20.2% who did so in 2012 and added up to only 90,000 votes versus Trump’s 180,000 vote margin of victory.10

Online post by Mark Robinson, who struggled with subsequent accusations of antisemitism in his North Carolina gubernatorial contest with Jewish candidate Josh Stein.

While Georgia had no statewide races, North Carolina had many, electing a mixture of Democrats and Republicans. The best performing Democrat had a net margin 18% better than Kamala Harris: Josh Stein, the first Jewish gubernatorial candidate in North Carolina’s history, won by a margin of of 14.80%, a lopsided result partially fueled by widespread perception that his opponent, Mark Robinson was antisemitic.11

Nevada & Arizona

Similarly, there are no indications the Gaza protest movement was significant in southwestern battleground states. Kamala Harris lost Arizona by 5.51% & Nevada by 3.10%, with Trump winning an outright majorities. “None of These Candidates” had earned only 7,448 votes in the Nevada primary.

Jackie Rosen is two to the right from Amy Klobuchar in this photo in that same delegation of senators to Israel previously mentioned.

Nevada also re-elected a staunchly pro-Israel Democrat, Jackie Rosen, who ran almost 5 points ahead of Kamala Harris, winning by 1.65%. In addition to being endorsed by AIPAC and with a track record of clear support for Israel, Jackie Rosen was one of three prominent Jewish candidates who outperformed Kamala Harris in a key battleground state. The winning Democratic candidate for Arizona’s senate seat, Ruben Gallego, faced protests against his support for Israel; he supported a ground invasion of Gaza following the October 7th attacks and has been criticized by the social justice left for his positive relationship with AIPAC.

What do the polls say?

In general, exit polls indicated that, as is usually the case, very few voters identified foreign policy as their primary motive for voting, with most exit polls pointing to 4%. In a close election like 2024, this could make up the difference if swing voters were more likely to support Palestinians than Israelis. Except there’s no reason to believe this is the case. Even 4% of the population is much larger than the size of the Gaza protest vote.12

Blueprint - swing voters who chose Trump on the problems with Democrats on Israel & Palestine.

Blueprint has carried out some interesting issue-oriented polls. They suggest that swing voters generally did not believe Democrats supported Israel too much, and generally did believe Democrats were too sympathetic to the Palestinian cause - a result that is in line with the fact that pro-Israel Democrats generally outperformed Harris while anti-Israel Democrats generally underperformed Harris.13

The Gaza protest movement was smaller than it looked

The size of the Gaza protest movement and its ability to mobilize voters in the presidential election was vastly overstated in media coverage. This was due to several factors. I already mentioned that the “Uncommitted” movement’s level of support was negligible outside of a few very specific areas, such as Dearborn and Minneapolis.14

The protest movement was overrepresented on social media relative to real life, and physically concentrated tightly within three areas that are disproportionately populated with social media influencers and traditional media figures:

Major cities.
Elite college campuses.
Areas with significant international presence.

Pro-Palestinian protests at Columbia, located in New York City. At its height, the encampment protest included dozens of tents and hundreds of students, a small albeit very visible fraction of Columbia’s student body.

The most visible protestors were students at Columbia, an elite college in New York City - the alma mater of many prominent journalists and right next to where many major media companies are. “Blue” media figures also tended to come from an ideological perspective that was more sympathetic to Palestinians.

A significant number of highly-involved protestors (particularly protest organizers) had close ties to foreign countries, and many were noncitizens.15 Contrary to popular myth, noncitizens generally don’t vote.16

A final recap

It’s clear that Kamala Harris did not lose because of the Gaza protest vote. It’s also clear that anything that Kamala Harris did to try to accommodate the Gaza protest vote would have come at an additional political cost.

Only Michigan and Wisconsin had a margin close enough to theoretically have been tipped by the Gaza protest vote, and winning those states would not, by itself, have salvaged Gaza.
It is not likely that the protest vote was large enough match Kamala Harris’s margin of loss in any state even in Michigan and Wisconsin.17
Kamala Harris would have lost other votes to Trump if she had been vocally pro-Palestinian.
Across the national political battleground, Kamala Harris consistently ran behind candidates who were vocally pro-Israel, Jewish, or both.18
Issue-based polls showed Democrats losing ground for siding more with Palestinians over Israel.
The Gaza protest movement was deeply unpopular among mainstream Americans - it just happened to be very visible on social media and traditional media.

If Kamala Harris made a strategic mistake on Israel and Palestine, it was in failing to distance herself from the pro-Palestinian wing of the Democratic Party - a pro-Palestinian pivot would have deepened her loss.

This favorability does not extend to Netanyahu in particular, but does extend to the Arab-Israeli conflict. For example, “sides with Israel over Palestinians” is at +24% in that survey, having fluctuated in a range from +22% to +48% over time. Given that recent polling shows bias in favor of “highly online” opinions, polls are more likely to understate this gap than overstate it.

Anti-Israel views are particularly common among immigrants and temporary residents because international opinion of Israel is more negative than American opinion of Israel, but these are not voters.

There were 45,000 “Uncommitted” votes in total. My best is that about 20,000-30,000 of those were actual protest votes, since it’s normal for a significant minority of voters to select “Uncommitted” in a Democratic primary.

It’s worth noting this difference was smallest in Ilhan Omar’s district where anti-Israel voters are most strongly concentrated, although Klobuchar outperformed both Harris and Omar in MN-05.

3.8% of the Democratic primary electorate voted “Uncommitted” in 1996. All things considered, it’s likely that the “Uncommitted” movement actually reduced the total number of uncommitted ballots nationally by branding the choice as being an act in their support. It seems very unlikely that any more than 1% of the Democratic primary electorate cast ballots in support of the Gaza protest movement, with 0.5% being a much more reasonable estimate - an estimate that is only that high because of the New Hampshire primary’s “Ceasefire” write-in and the localized pockets associated with the congressional districts represented by Rashida Tlaib and Ilhan Omar.

Technically “Uninstructed” in the case of Wisconsin.

This includes Jill Stein and Cornel West. It does not include voters who left the top of the ballot blank (top-ballot undervoting is rare) but also includes the small number of perennial Green voters who were never going to vote for a Democratic candidate anyway.

Both also saw improvements in their approval ratings in polling conducted in the wake of their public criticism of the Gaza protest movement.

Kenyatta, who lost by 5.6%, largely avoided commenting publicly on Israel this cycle other than a very straightforward condemnation of the October 7th attacks, but has previously expressed normal party-line sentiments on Israel back when he was running for Senate. (He lost the primary to Fetterman.) As a first estimate, it would not be inappropriate to say that “viewed as hard social justice left” was around eight times more important in terms of lost votes than “supports the BDS movement.”

As one of the single largest contributors to the “Uncommitted” movement’s claimed national tally, this underlines the fact that the tally of uncommitted votes vastly outran the movement’s actual base of support.

This was reasonable considering Robinson’s many controversial statements. He engaged in Holocaust denial, called himself a “black Nazi” while commenting on a pornographic internet forum, and at one point stated that he would have preferred Adolf Hitler as president over Barack Obama.

Jill Stein earned 0.5% of the popular vote overall, and likely attracted (as she did in Dearborn) a majority of the actual protest votes.

This was not limited to battleground states. Bernie Sanders, usually a strong performer, has been highly critical of Israel in the last year and underperformed in relative to Harris. Elizabeth Warren has also been highly critical of Israel, but has an existing history of electoral underperformance.

I.e., probably as much as half of the uncommitted vote in Michigan and Minnesota was not cast on behalf of the “Uncommitted” movement. That share of unrelated uncommitted votes could easily have exceeded 90% in North Carolina and other states not home to clear visible pockets of support.

This was given as a reason why MIT did not suspend many of its student protestors in 2023: As they were on student visas, suspensions would have led to deportations.

In addition to the very low rates of detection of noncitizens attempting to vote, states with high immigrant populations have correspondingly lower populations of registered voters and actual voters. Indeed, even naturalized citizens tend to have low voter turnout rates.

As I pointed out, the protest vote in the general election was more plausibly approximated by the Jill Stein vote than the uncommitted primary votes.

To recap, candidates with stronger pro-Israel credentials outperformed Harris in Minnesota (which she won narrowly), Michigan, Nevada, Arizona, Pennsylvania, and ME-02, while Jewish candidates outran Harris in Michigan, Nevada, and North Carolina. Candidates with stronger pro-Israel credentials were not present on the ballot in Georgia, New Hampshire, or Wisconsin.

I didn't beat the polls

Tomas McIntee — Sun, 24 Nov 2024 17:55:50 GMT

When I presented a poll-agnostic historical model for election forecasts, I promised to compare that model to the polls after the election.1 So, here we are: I’m going to drill down into polling accuracy: How bad was the miss? And how weird, exactly, were the polls?

Many wrong takes were published about the popular vote in the short period immediately after the election. However, while it’s common practice for news outlets to “call” elections promptly, some states count ballots slowly for various reasons.2 This article couldn’t really be written before waiting for the counting to reach a near-complete state.

The topline results

Overall, the results were more strongly pro-Trump than 2016. Since the trend 2016-2020 was in a Democratic direction, this meant that the full range of historical models was wrong, with the closest one overall being the average model. By state, the average error of the margins in that model was 2.9% (underestimating Trump) and the RMSE was 4.3%. Within the seven key battleground states, the RMSE was lower, only 2.9%.

My range of historical models published before the election

By comparison, the mean error by state in the FiveThirtyEight median projection was only 2.7%, with the RMSE by state being 3.3%, sharpening to 2.5% for the seven key battleground states. SplitTicket’s projection missed very similarly (3.5% RMSE, 2.5% in battleground states). What this means is that I didn’t beat the polls. There was some information there.

That said, most of the difference between the historical projection and the poll-based projections came from results in non-battleground states. Pollsters who chose to poll in non-battleground states picked up signs that Trump was improving from his historical baseline in New York, California, Florida, and Massachusetts.

Polls didn’t do exceptionally well or poorly, but they missed, mostly in one direction, and I think it’s worth exploring why.

The switch was a bad bet

When I published an article announcing that I expected that poll-based models wouldn’t do any better than purely historical models, I said that if Kamala Harris outperformed the historical models, it would mean she was a good candidate; and if she managed to underperform them, it would mean she was a bad candidate. Historically, voters have responded poorly to Democratic elites overtly selecting the Democratic nominee; c.f. 1968 & the response to the DNC e-mail scandal in 2016.

Harris underperformed the average of 2016 and 2020 both nationally and in 41 out of 50 states. What this suggests, fairly directly and pointedly, is that the switch from Biden to Harris was a bad idea. The counterfactual claim that Biden would have done worse lacks support. While Biden’s polling numbers were not good, looking closely at the polling numbers suggests that Harris may have been the beneficiary of stronger polling bias rather than stronger underlying support.

Biden versus Harris: The polling oddities

To start with, it’s worth noting that political polling was already unusual when Biden was still in the race. It’s also worth noting that

While his approval ratings were historically low, this was driven by “soft” within-party disapproval, and was not reflected in actual voting behavior in the 2024 presidential primaries. In contrast to polling and the expectations of vocal Biden skeptics on social media, neither Dean Phillips nor Marianne Williamson nor the pro-Palestinian protest vote gained significant support.3

As I noted previously, polls dramatically overestimated the opposition to Biden in the 2024 presidential primaries, and dramatically underestimated the opposition to Trump in the 2024 presidential primaries. The fact that polling was biased in favor of Trump relative to Biden but in favor of Harris relative to Trump is difficult to explain.

The same pattern, interestingly, was visible in New York Times headlines covering primary results after the fact.4 Overall, Trump won 76% of the primary vote and Biden 87%, but both media coverage and polling were in line with an alternative reality where Trump won 87% of the primary vote and Biden only 76%. Polling was, in other words, biased in the same direction as what Democratic media elites seemed to want to believe.5

This extended to very particular details, such as the sudden Harris lovefest that occurred among Democratic media elites after Biden stepped back and endorsed her:

I don’t think this sudden discontinuity reflected any sudden improvement in Kamala Harris’s public standing as far as future voting behavior was concerned. Instead, when Kamala Harris became the presumptive nominee, loyal partisan Democrats approved of Kamala Harris being the nominee and rallied around the flag to support the new presumptive nominee.

Which brings me to the next interesting polling pattern. What happens when we look at opinions of Harris outside of loyal Democratic partisans?

Plot of FiveThirtyEight polling averages for Trump’s share of the vote. The Silver Bulletin

Accompanying the sudden jump in Kamala Harris’s favorable ratings as loyal partisan Democrats rallied around the flag was a jump in Donald Trump’s head-to-head polling numbers. This was part of a long-running pattern visible in head-to-head polling back in 2019: Based on polling, a small but notable segment of the population was willing to vote for Trump to Harris even though they didn’t like Trump and would otherwise vote for a Democratic candidate.6

This sudden jump in Trump’s average was not just an artifact of 538’s algorithm, nor did it occur in response to the first debate,7 nor does it date from RFK’s withdrawal a month after the switch:

Plot of polling average and linear trend before and after candidate switch. Note that if we separate the period between the debate and Biden dropping out, the trend in Trump support in that period is actually negative.

This is the median voter theorem at work. When Kamala Harris became the presumed standard-bearer, Democrats began to shed moderate swing voters, beginning first with highly-informed swing voters who were already familiar with Harris’s progressive record and continuing as voters became more familiar with Harris nationally.

Democratic media elites never wanted Biden as a presidential candidate in the first place and continued to lobby for the replacement of Biden as nominee.8 They could have backed a primary challenger; they didn’t. They could have accepted that Biden’s debate night performance was an unusual event, the result of exhaustion and illness; he held an energetic rally in North Carolina the next day and gave a lively master class in foreign policy to a hostile press on July 11th.

Why did polling miss?

Since baseline response rates are extremely low, polls are highly vulnerable to selection effects. My conjecture is that the following five issues are broadly problems for pollsters:

Polls oversample high-mobility people.9
Polls oversample people with significant free time during regular work hours.
Cell phone polls undersample people who change cell phone numbers more often.
Online or partially online polls undersample people who have more limited internet access.
It is difficult to poll “double haters,” who now make up a large and significant share of swing voters.10

The first four factors on this list lead to a systematic oversampling of the highly online group. This skews towards highly educated people, but also includes people in the same social groups who don’t personally have impressive credentials. Interestingly, this error means that polling would be skewed in the same ways that opinion among blue media elites was skewed.

That is to say in favor of Harris over Trump…

This is a very important graph, so I’m posting it twice. Because polling underrepresented the opinions the groups Harris did poorly with, this graph can be expected to understate the degree to which Trump’s support increased after Harris became nominee.

…but it also means that polls were biased in favor of Biden over Harris. Both polling and primary election results generally indicated that Biden was stronger than Harris among minority voters and blue collar voters, who have been effectively underrepresented in polls, while Harris was most popular among highly-educated white voters.11

Actually, a parameterized range of models, based only on the previous two elections.

Two notable reasons: Absentee ballots that arrive after Election Day with postmarks from before Election Day and provisional ballots (ballots cast by voters who could not prove eligibility properly on the spot). It’s not just incompetence on the part of state and local officials.

With 87% of the total popular vote, Biden performed very similarly to Obama in 2012 (90%), Clinton in 1996 (89%), rather than Carter in 1980 (51%).

The main episodes of low support for Biden were in New Hampshire (where he was a write-in candidate not listed on the ballot), Ilhan Omar’s congressional district, and Rashida Tlaib’s congressional district.

Both polling and Democratic media elites frequently described the opposition to Biden among Democratic voters as more significant that the opposition to Trump, in spite of the fact that Trump had almost twice as much dissent with 76% of the vote.

Either NYT editors didn’t think it was worth revising already-written articles based on expectations that were proven false, or the people writing the headlines for articles simply wanted to present Trump as an electoral juggernaut and Biden as struggling in spite of the sharp numeric contrasts.

This should not be interpreted as anything nefarious, but it’s probably not entirely coincidental, either.

I.e., Trump won the “double haters” per exit polls.

If we break up the regression to have a third piecewise segment between the debate and the speech, that segment shows Trump’s support dropping. “Before-and-after” panels of voters showed little response to the debate, in contrast to the very dramatic on-the-spot media reaction.

There was a slow and steady trickle of two types of articles in Democratic-leaning media outlets, with the New York Times being particularly obviously slanted against Biden. This included normal opinion articles and also persistent false claims that Biden had pledged to only serve one term during his 2020 campaign, based either on “anonymous sources” or on close readings of ambiguous political rhetoric.

These claims started up in 2019 before Biden was the Democratic nominee and were consistently met with unequivocal public denials from the Biden campaign.

For example, after moving back to North Carolina, I continued to get political pollsters calling me about races in California, Virginia, and Ohio, because I was on lists for areas I no longer lived in.

Historically, it was typical for both presidential candidates to have positive net approval from the public. The only candidate in the Trump era to have positive net approval for any part of their presidential campaign was Joe Biden in 2020, and Joe Biden’s net positive approval ratings were not very strongly positive by historical standards.

I.e., even when minority and working class voters are reached in polls, they tend to be minority and working class voters with opinions less representative of their groups, skewing polling results away from actual minority and working class voting behavior.

How did we get here?

Tomas McIntee — Sat, 16 Nov 2024 19:37:21 GMT

The last time that Democrats or Republicans nominated a candidate that was viewed favorably by a majority of Americans was in 2012. Out of six major party nominations, five have gone to candidates with negative net approval ratings (the lone exception being Joe Biden in 2020), and five have gone to geriatric candidates (the lone exception being the irregular nomination of Kamala Harris this year).

The four winners of the last six major party presidential nominations. Note that after getting into office, there were brief periods of “honeymoon” positive net ratings (~1 month for Trump as president, ~1 year for Harris as vice president). Biden had positive net approval ratings until he ordered the withdrawal of American troops from Afghanistan, but was the oldest of all four presidential candidates.

Over the last ten years, both parties have repeatedly failed the American public. The cause? In short, media. Changes in traditional media and the rise social media have created a divided and highly partisan collection of information bubbles, isolating energetic partisans from each other and isolating elite decision-makers from ordinary Americans.

It would be easy to say that the problem is Trump, but these problems have become rife in both parties from top to bottom.

Yes, 2024 election showed weakness by Republicans

When I initially drafted this article in October, I expected Trump to win 46-47% of the vote as he did in 2016 and 2020; at present, it seems as if Trump has won 50% of the popular vote. I knew that Kamala Harris was a particularly bad candidate, but I expected the variance to come out in third party protest votes.

While Republicans can (and should) claim victory in the 2024 election, there are some important warning signs that Republicans should heed:

Trump remains persistently unpopular.
Republicans lost the most prominent statewide race on the ballot in five out of the seven closest battleground states.1
Republicans remain on the losing side of several important issues.2
The Republican part of the Horseshoe Caucus is larger and more prominent.

2024 Senate elections. Republicans picked up a net three seats, but still lost in Arizona, Michigan, Nevada, and Wisconsin; all four states were won by Donald Trump.

While I find it hard to continue to justify my pre-election opinion that Nikki Haley would obviously do better in the 2024 contest, as Trump’s performance exceeded my expectations, it is important to keep in mind that this was not, by any standard, a blowout election: Trump’s popular vote margin was slimmer than the winner in ten of the last twelve elections.3 Trump won by a narrow margin against one of the three least popular Democratic nominees in living memory.4 The fact that downballot Republicans underperformed Trump when Trump was less than 2 points ahead of Kamala Harris nationally should be alarming to Republicans.

Trump isn’t the only problem

Trump has become the defining visible icon of political dysfunction 2016-2024. Trump has been an incredibly unpopular political figure, both in his three election cycles as a candidate and during the four years of his presidency.5 Trump’s unprecedented level of success as a political figure with consistently negative levels of popularity is not the root cause of political dysfunction, but a consequence of political dysfunction.

The fact that Donald Trump won the presidential election of 2016 while losing the popular vote 48.2% to 46.1% was an anomaly of the Electoral College. The Republican Party nominating Trump was a colossal failure that should have (and nearly did) result in the loss of the presidential election.

Subscribe now

The fact that Republicans held control of Congress in 2016, however, was not a stroke of unusual luck, nor was it due to gerrymandering.6 To put it bluntly, Democrats had a losing political strategy in 2016, and only the unwise nomination of Donald Trump gave Democrats any reason to hope for victory.

How Democrats lost in 2016

In 2016, Hillary Clinton won the popular vote nationally amid high rates of overall third-party voting, but lost the popular vote in six hotly-contested states that Barack Obama won in 2012: Florida, Pennsylvania, Ohio, Michigan, Wisconsin, and Iowa. Four of those states had a Senate contest in 2016; Democrats lost all four of those contests. Democratic candidates won fewer votes for the House in five of those six states.7 The fact that Hillary Clinton won the popular vote against Donald Trump concealed the fact that Democrats had embraced losing political positions, including:

Clinging to an unpopular neoliberal consensus on trade and immigration.
Veering left on numerous other social and cultural issues.
Attacking voters as racist and sexist for supporting the wrong candidate.8

How unpopular was Clinton? A septuagenarian socialist came in from outside the party and gave Clinton the second-closest Democratic primary contest since the McGovern-Fraser reforms. (Image source.)

The national popular vote for the House of Representatives is one of the best proxies for a national popular vote outside of the presidential election. In a year where Donald Trump lost the popular vote by 3 million votes, Republican candidates for the House of Representatives won approximately 2 million more votes.9 This is not because Republicans ran a larger number of candidates in uncontested seats.10 Nor can it be blamed on gerrymandering.11

Overall, on the national scale, Hillary Clinton won almost exactly the same share of the vote as downballot Democrats. She only won the national popular vote because swing voters chose third party candidates over Donald Trump: Her campaign was no more and no less than the face of the 2016 Democratic Party, with an unpopular platform and unpopular tactics.

How did Democrats veer off-course?

If we look back, Democrats arguably had the winning portfolio of issues continuously from 1992 to 2012 - despite significant losses in lower-turnout midterm elections in 1994 and 2010 and a brief (2002-2004) surge in Republican popularity linked very specifically to the 9/11 attacks and the subsequent War on Terror.

Image from source.

Around 2012, things changed. One key factor in the timing of this shift was the widespread adoption of smartphones and social media apps. Online left-wing and right-wing political communities rapidly became both larger and more insular, elevating more extreme politics as a form of social group identity. This directly affected many influential Democrats, including political leaders, activists, and media figures.

Democrats shifted on social and cultural issues

For Democrats, the major shifts associated with the rise of social media were the rapid widespread adoption of what had previously been considered fringe positions on social and cultural issues. “Cultural issues” in general were identified in at least one prominent issue-based exit poll as the top tested issue for swing voters who decided to vote for Trump, with immigration concerns identified as the third and fourth most important issues.12

Political cartoon by prominent trans dissident Colin Wright, drawn to explain how he came to be considered right-wing.

Thus, while most Americans found the idea of “safe, legal, and rare” abortion appealing, Democrats struck the word “rare” from their platform in 2012 and began to use unqualified support for abortion as a litmus test for candidates. Overall public opinion on abortion, on the other hand, remained largely static from Roe v. Wade (1973) to Dobbs v. Jackson (2022); with the notable specific exception of gay marriage, the median voter’s opinion shifted on very few social issues from 2008 to 2016.13

Share MathIntee Political Post

Other key issues on which 2012 marked a turning point for the Democratic Party included gay marriage and the rhetoric surrounding immigration.14 Highly-online Democrats and more broadly the highly-educated upper class Democrats who shared social circles with highly-online Democrats continued to rapidly shift left on social issues from 2012 to 2016, adopting more and more of the languages and principles from the “social justice” wing15 of the growing online political community.16

Republicans abandoned the neoliberal consensus first

The 1976 Democratic Party platform savaged the Nixon administration for failing to “effectively resist” the “unfair trade practices” of the Soviet Union in 1976 while expressing carefully qualified and limited support for the principles of free trade. By 1992, however, many Democrats were as strongly in favor of free trade as Nixon. There was a neoliberal economic consensus.17

Both parties catered to business interests that wanted unfettered access to foreign markets and cheap labor. Arguably, this neoliberal consensus began to form over immigration policy in the 1960s, with shared unpopular positions that could generously be described as nuanced and less generously as muddled.18 From 1968 to 2008, little difference could be discerned between the platforms and policies of both parties on immigration, with a separation beginning with what could have been considered merely rhetorical divisions in 2012.19

Image source.

From 1992 to 2012, the median voter was more protectionist and nativist than either major party. The opposition to the neoliberal consensus cut across partisan divides, with the organized manifestations of this opposition relegated to left and right fringes. Within the Democratic Party of 2016, this opposition aligned with Bernie Sanders and was largely marginalized by the victorious Hillary Clinton camp.20

The online right wing, however, embraced this skepticism without qualification; in 2012, the Republican platform brought up concerns about growing unfair trade. In 2016, these concerns turned into full-throated criticism of massive trade deficits and a claim that free trade had not been fair trade. The 2016 cycle also saw sharp shifts in the Republican platform on immigration.

The median voter reigns supreme.

By 2016, Republicans had abandoned the unpopular neoliberal consensus, while Democrats had shifted sharply to the left on a number of social issues, leaving the median voter behind. This shift was visible in 2012 in the activist wings of both parties and among party elites by 2016.21

While there were many reasons for Clinton’s loss in 2016,22 the main reason why the race between Clinton and Trump was close enough for these small factors to matter was the main reason that Democrats lost control of Congress: They moved away from the median voter and were punished accordingly. Democratic elites picked a weak candidate and, under the leadership of that weak candidate, built an unappealing platform.23

What happened in 2020?

Following Hillary Clinton’s loss in 2016, the rapid leftward movement accelerated. True believers laid out the case that Hillary Clinton had been too moderate, and that this had depressed enthusiastic turnout. Bernie Sanders pivoted to the left on immigration and social justice issues and moved further left on economics; most 2020 presidential primary contenders followed suit.

Unlike in 2016, however, party elites and media elites were not on the same page. Both party and media elites disliked the true outsiders (Sanders, Gabbard, and Yang), but Democratic media elites were additionally hostile to candidates who were either moderates or white males, with the partial exception of Pete Buttigieg, the gay mayor of South Bend.

Differences in media coverage relative to polling numbers. Note that Bill de Blasio received substantial coverage related to being mayor of New York City, while Jay Inslee received substantial COVID-19 coverage related to being governor of Washington. Swalwell’s brief surge of coverage during his very brief campaign came a highly-publicized attack on Biden at one of the early debates.

The field of potential contenders was extremely large. In helping winnow the pool of candidates with their selective attention, “blue” media outlets exhibited strong and visible biases both in terms of the volume and tone of coverage as Democratic media elites tried their hardest to ensure that the nominee was neither a white male nor a moderate. This meant pushing against more electable candidates in favor of less electable candidates.

However, while Democratic media elites were vigorously opposed to the nomination of Joe Biden, none of the media favorite candidates were electable candidates, and party elites were comfortable with backing Biden. Most Democratic voters wanted a qualified electable moderate.24 Like Sanders, Biden was already arguably too old for the job, but younger moderates lacked the national prominence of Biden and Sanders. Thus, they were easily shoved aside by a media ecosystem with unlimited hostility for moderate Democrats.25

With Biden at the top of the ticket and the moderate wing in charge, Democrats also abandoned the neoliberal consensus, returning to a more traditional “labor left” approach, blasting a history of unfair trade and outsourcing; the phrase “free trade” only occurs in the 2020 platform once, in a reference to the African Continental Free Trade Agreement.26

The 2024 coda

While a full discussion of the Biden administration is beyond the scope of this article, I will note a few key points. Biden adopted a “labor left” approach to economics that expanded on Trump’s tariffs,27 but decided to have his vice president, Kamala Harris, lead the effort to reduce the flow of illegal immigrants across the border. Instead of decreasing, the flow sharply increased.

In 2024, in spite of serious challenges from within the party, Donald Trump was renominated to run as the Republican candidate for a third time. Joe Biden became presumptive nominee with no serious challengers until he resigned from the ticket. Loud vocal disapproval from within progressives and blue media elites fueled negative net approval ratings that were nearly as low as Trump’s, and polling suggested that large numbers of 2020 Biden voters were undecided.28

Abandoning the presumptive nominee is risky.

After the primaries had concluded and an early debate took place between Biden and Trump, Biden resigned from the ticket.29 This process was initiated by media elites who had always preferred other candidates to Biden all along (beginning in 2019) and picked up support from a few key Democratic insiders. The Democratic nomination was then passed, logically enough, to Biden’s running mate Kamala Harris, a move that was a risky bet, especially considering past precedent.

Before becoming Biden’s vice president, Harris had run in the progressive lane. She was a bad candidate in 2024. Thanks to her record, it was not difficult for the Trump campaign to convince swing voters that Kamala Harris was a far-left “social justice” candidate who represented everything unpopular about the Democratic party.30 She was not unburdened by what has been, and neither voters nor Republican advertisements treated her as if she had fallen out of a coconut tree the moment she was endorsed as by Biden as his replacement.31

It isn’t just about weak parties

Some political scientists have argued that much of the current trouble with America’s two major parties relates to the parties being “weak.” That is to say, party officials and party insiders have limited control over the nomination process, which is a party primary vote. This allows insurgent candidates like Donald Trump to effectively take over a party by popular force. It is true that a strong Republican Party would not have chosen Donald Trump, because Trump did not represent the ideals of the party and was deeply unpopular among party insiders. In addition, he was too old.32

However, across first world countries, the exact same dynamic has played out very similarly with “strong” and “weak” party systems, leading to the rise of far right parties across Europe. The core problems relate to a rapid shift in elite opinion (including media elites as well as traditional party elites), linked closely to the rise of social media.

European parliament in 2009 (left) vs 2024 (right). Here, the left wing is in red, the right wing is in blue, and the new hard right parties are dark blue. Note that the far left also made modest gains, similar to the rise of the DSA & Justice Democrats within the US Democratic Party.

Even within the United States, Republicans are not the only party that has failed the American public in the last three cycles. Hillary Clinton was the selection of the large majority of Democratic insiders prior to the 2016 primaries. Democratic elites selected the least electable candidate out of the available options and then led the way in steering their platform away from the median voter.

In the 2020 cycle, Biden was too old. However, Democratic media elites were successful in preventing the rise of younger “moderate lane” candidates; party elites then rallied behind Biden because he seemed to be the only remaining candidate able to defeat Sanders in the primary election and Trump in the general election.

Kamala Harris was in second place in the “invisible primary” of the 2020 cycle until her campaign imploded, and was able to gain support across a range of party and media elites rapidly during the anti-Biden pressure campaign. Even if Biden had stepped aside earlier, Kamala Harris was the “next in line” insider candidate and highly capable at “inside track” politics, precisely the sort of candidate favored by “strong” parties.

The rise and reflection of bigotry

Social media polarization and the partisan capture of traditional media has also led to a steep rise in the acceptable levels of bigotry within both parties. Large segments of the “blue” media environment tolerate discrimination against men, whites, Asians, Christians, and Jews; similarly, on the “red” side, there are safe places for bigotry against women, blacks, Hispanics, Muslims, and Jews.33

Meanwhile, most Americans, and especially most swing voters, have a limited tolerance for bigotry. In 2016, this probably hurt Donald Trump at the polls. However, by 2024, the bigotry of the left fork of the political horseshoe was on clear display, notably in the Ivy League and local city governments.

“The Boy Who Cried Wolf,” two separate public domain illustrations.

During the first Trump administration, critical race theory and other fringe left-wing movements gained significant public traction. Accusations of racism flooded out from the left side of the political spectrum; everything from mathematics education to incarceration was racist. The criminal justice aspect of this was particularly visible, with demands for the defunding of police, the abolition of prisons, and de facto decriminalization of petty crime.

This movement achieved peak levels of success from 2018-2022 and had a significant effect on how “blue” cities were governed during this period. By 2023-2024, it was clear that the experiment had failed.34 This not only discredited the politicians who had embraced these policies, but also blunted the credibility of the left in making accusations of racism.

And then the Ivy League problems hit. 2023 was the year that the Ivy League became widely known to moderate American voters as racist, and this reflected poorly on Democrats, as the Ivy League is no less “deep blue” Democratic than inner cities.

It started with a Supreme Court case, SFFA v. Harvard. It was already well-known that the Ivy League had been widely discriminating against Asian-American applicants and in favor of African-American applicants, but the case brought this practice to public attention. The Biden administration filed an amicus brief, and many Democrats spoke out in favor of affirmative action, but affirmative action is incredibly unpopular among the American public.35

The Supreme Court decided 6-3 (as pictured, minus the physical distance) that affirmative action programs in college admissions were racist and illegal.

Several months later, Gazan militants invaded Israel, murdering over a thousand people and taking hundreds of hostages, including dozens of Americans. Ivy League campuses immediately erupted with pro-Palestinian demonstrations, including celebrations of the attack, preemptive protests against Israel’s response, and genocidal chants calling for erasing Israel from the map.36 While not all of the protesters were antisemitic, antisemitism was very clearly tolerated and present.

Ivy League university presidents were summoned to Congress to testify about why they tolerated hate speech and an environment hostile to Jewish students. This, in turn, inspired the investigation of Harvard University’s president, Claudine Gay, whose scholarly record turned out to be rife with obvious plagiarism. This, in turn, suggested her work (and her resume) had been subjected to lower standards because of her race and sex… which made for the third distinct major news cycle in four months that focused attention on bigotry on Ivy League campuses.

What happened?

To summarize, we’ve entered an era where social media has assisted a rapid polarization among activists and ideologues. This has accelerated the rise of highly-partisan media outlets. The result is that both parties have embraced deeply unpopular candidates that are popular among their respective activist classes, with party elites that are highly responsive to partisan media elites.

Both parties have growing problems with bigotry, which is repugnant and repulsive. Unfortunately for Democrats, “blue” bigotry was in the spotlight for much of 2023 and 2024.

While Republicans arguably were “captured” by partisan media first and selected a standard bearer who was more deeply unpopular on a personal level, Republicans were first to abandon the neoliberal consensus and Democratic elites have shifted sharply away from the median voter. This shift has been especially pronounced with Democratic media elites, who appear to be pulling traditional Democratic party elites in their wake.

This includes Senate races in Arizona, Michigan, Nevada, and Wisconsin, as well as the gubernatorial race in North Carolina.

Abortion and the war in Ukraine. Donald Trump

The exceptions are the 2000 and 2016 elections, where the winner of the Electoral College did not win the popular vote at all.

At the time that Joe Biden dropped out and Kamala Harris became the presumptive nominee, her average net approval was 13 points underwater and had been underwater for almost three years. Hillary Clinton’s average net approval bottomed out at 19 points underwater, and had only been negative for about a year before her nomination; Harry Truman’s net approval bottomed out at around 14 points underwater, rebounding to a mere 4 points before the infamous 1948 Democratic National Convention.

Donald Trump has had negative net favorability for almost his entire career as a public figure. While polls have slightly underestimated his level of popular support, his net approval ratings have generally exceeded the magnitude of observed polling bias. This includes exit polls from when he won the popular vote in 2024, where his net approval ranged from -7% to -11% (1,2,3); Trump did not win in 2024 by becoming better liked, but by earning the votes of “double haters” who disliked both Trump and Harris.

The fact that Donald Trump enjoyed net positive job approval for less than one month out of his entire presidency is completely unprecedented in the history of modern polling. All things considered, there are only a handful of presidents in history that could plausibly rival Donald Trump for persistent unpopularity in office. This includes John Quincy Adams (elected in spite of placing second in the Electoral College), James Buchanan (elected just before the Civil War in a period of extraordinary political dysfunction), and various unpopular vice presidential successors (most notably Andrew Johnson).

Gerrymandering created a Republican House majority with a minority of the vote in 2012, but in every subsequent election, the party winning more votes overall has won more seats in the House.

In Wisconsin, the lone exception, there were two uncontested seats. Granted, those seats were strongly Democratic, so the total popular vote would have been expected to be close. There was also an uncontested Republican seat in Pennsylvania, which most likely would have narrowed but not eliminated the Republican popular vote victory at the House level.

This started during the primary season with widespread attacks on “Bernie bros.”

There are several caveats involved with coming up a precise total.

From 2012 to the present, there are generally more uncontested Democratic seats, for reasons mostly related to California’s top-two system.

Gerrymandering strongly impacts the translation of votes into seats, but does not strongly impact vote totals for either party except via the mechanism of uncontested seats.

Inflation was second, with all three top issues being very close in importance.

Note that this has not been a blanket shift on all LGBT issues. Particular areas in which online left community consensus has rapidly moved away from the median voter include a large bundle of trans issues (transwomen in women’s sports, medical transition for minors, legal recognition for social transition) as well as the conflicts between religious freedom and anti-discrimination measures.

The 2012 platform was the first to use “LGBT” and support gay marriage. On immigration - the 2008 Democratic platform stated that “We cannot continue to allow people to enter the United States undetected, undocumented, and unchecked.” The 2012 platform contained no such language, with the DACA program representing a monumental shift in the Democratic approach toward immigration.

There was also left-wing online communities that had a strong focus on economic issues over social issues, but that faction lost, both in terms of online community networking and in terms of influencing the Democratic Party in 2016; they backed Sanders in the 2016 primary.

The 2016 Democratic platform used a variety of newly-popularized phrases such as “LGBT elders,” “identify as,” “restorative justice,” “racial justice,” “climate justice,” “environmental justice,” and “health equity.”

Fully exploring this topic really deserves its own article; the neoliberal consensus begins with the field of economics.

The linkage here is cheap labor; the growth of illegal immigration was linked to widespread labor law violations. While it might seem difficult to believe now, the 2012 Republican platform celebrated the fact that one million immigrants were granted permanent residence per year and called for increases in legal immigration, particularly for high-skilled workers, while rhetorically condemning illegal immigration in terms very similar to the 2008 Democratic platform.

DACA is the only notable de facto policy shift, but was the conclusion of a substantial bipartisan effort (with similarly bipartisan opposition) linked to the DREAM Act, beginning in 2001. A close parallel can be drawn between DACA and Reagan’s immigration amnesty program operating from 1987 to 1988, which similarly focused on minors and had a mixture of support and opposition that similarly cut across party lines.

Frequently overlooked is the fact that Bernie Sanders in 2016 had a record strongly skeptical of immigration as well as global free trade. Sanders and Sanders allies subsequently pivoted on this issue in order to try to win Democratic primaries in 2018-2020. Skepticism of globalism was a major factor in crossover support for Sanders in 2016 as well as Ron Paul in 2008 and 2012.

The remarkable thing is that party elites didn’t seem to realize that they were adopting unpopular and radical positions.

I can (and have) pointed to a number of other factors, such as the Clinton campaign’s self-defeating prophecies of inevitable victory, the way that loyalist Democrats chose to smear Sanders supporters as sexist and racist, the backfire against the elites’ visible attempts to stack the deck against Sanders, the timing of James Comey’s announcement, etc. Trump’s margin of Electoral College victory was exceptionally narrow and fragile.

Bernie Sanders regularly performed better than Hillary Clinton in head-to-head polls against Donald Trump.

Some Democratic media elites warmed up to Klobuchar very late in the game as they grew desperate for any alternative to Biden and Sanders. Reception of Pete Buttigieg was mixed and sometimes favorable, but he lacked any of the traditional qualifications for a presidential candidate (congress, cabinet, governor, or general). After Klobuchar, the closest traditional “moderate lane” candidate was Beto O’Rourke, who was the single most under-covered candidate by the media relative to his polling numbers (and covered remarkably negatively given that he had been a media darling in 2018).

“Red” media also preferred to focus on “progressive lane” candidates, such as Elizabeth Warren and Kamala Harris.

Since Biden had been in office for more than a decade of the Democrats’ pre-consensus protectionist period, it would not be surprising to me if this was driven by Biden’s long-held beliefs as much as electoral opportunism.

Biden managed to lower unemployment while outpacing Europe on overall GDP growth, with inflation rates under the global inflation rate from 2022-2024 but still high compared to recent US inflation rates. Whether or not this was a “good” economic policy & how much credit/blame Biden deserved for the various parts of it was a mixed bag.

Within 2024 polling, it’s not clear if the overall difference was significant, it’s not clear whether or not this was significant. Head to head polling seemed to generally inflate Harris’s numbers over Trump’s, suggesting a general partisan skew; however, polling also substantially overestimated the opposition to Biden within the Democratic party during the primary season.

It is worth noting that Biden had higher ratings among independents and opposite party voters, but lower within-party approval than Trump, and that support for Trump in head-to-head polling remained at or below his actual 2020 vote share during this period; I’ll talk about this more in another article.

Contrary to what may be assumed from the media coverage, swing voter panels and polling indicated that the now-infamous debate between Joe Biden and Donald Trump had little to no impact on voter opinions or intentions.

On immigration, she was the failed “border czar” of the Biden administration, and as a presidential candidate had promised to decriminalize border crossings. On affirmative action, she was visibly an “affirmative action pick,” because Biden had publicly promised to pick a black woman as his running mate long before announcing Harris as his choice. She had one of the most progressive voting records in all of Congress during her time as a senator. On a wide range of social and cultural issues, she was already on the record as espousing “social justice” orthodoxy.

Many prominent “blue” media figures - both on social media and in traditional media - have suggested that Kamala Harris ran to the center as a moderate. This is partially true in that she tried to avoid saying anything new that would spook swing voters, but she also tried to avoid saying anything that would make progressive activists angry and ran what a lot of those same figures were happy to describe as a “vibes-based” campaign. This was not effective, although I am skeptical that, given her past record and quotable quotes, Harris could have won even if she had a clearly outlined strategy for trying to redefine herself as a true moderate.

I am firmly of the opinion that no major party should select a nominee older than William Henry Harrison for an open seat. This presents real risks.

The fact that both left and right wing extremist communities tend to feature antisemitism is one of the things that brings together the Horseshoe Caucus in opposition to Israel.

If it seems strange that it would take a long time for a critical mass of swing voters to identify these ideas as bad, remember that the COVID pandemic provided a major disruption to public life in 2020-2021, and that a number of major city governments had significant existing failures of governance involving bad schools, housing supply crunches, and high crime rates.

Remember, “left coast” California voters voted overwhelmingly to ban affirmative action in 1996.

E.g., “We don’t want no two state, we want all of forty-eight,” “From the river to the sea, Palestine will be Arab,” etc. See also “flood” marches, paraglider iconography, et cetera.

I told you so: Democrats made a bad bet

Tomas McIntee — Sat, 09 Nov 2024 22:31:13 GMT

This is not the long article laying out how things got to where they are today, or the series of continued failures by both major American political parties stretching back for a decade. That article will come, along with others. Hit the “subscribe” button to get them delivered into your e-mail box.

Today, I’m going to talk about something simple, something that is obvious to me from the numbers but clearly not obvious to everyone: Kamala Harris was a weak candidate, and the flaws in her candidacy and campaign were an important factor in why the Democrats lost.

Democrats could have bet on a repeat of 1948 or against a repeat of 1968.1 They chose the latter bet with Kamala Harris.

Kamala Harris didn’t just fall out of a coconut tree

One of the major problems with Kamala Harris as a candidate is her history. While not exceptionally progressive as California’s attorney general relative to California politics, her Senate record identified her one of the two farthest-left senators and three farthest-left members of the entire 116th Congress.2 She ran for the Democratic presidential nomination in 2019 in the “left” / “progressive” lane, which positioned her on the losing side of many of the issues that mattered to 2024 voters.3

In head-to-head polls conducted in 2019 featuring multiple Democratic candidates against Trump, Kamala Harris performed worse than Elizabeth Warren, Beto O’Rourke, Bernie Sanders, and Joe Biden.

Unsurprisingly, she had bad favorability polling and bad head-to-head polling. She even had a weak electoral record in California, which was not a battleground state at the time. All factors pointed to her being among the weakest candidates early on,4 even initially when the media coverage of her campaign was largely positive in tone.5 Five years ago, my best estimate was that Kamala Harris could be expected to underperform the typical Democratic candidate by 2.5%.6

Kamala Harris clearly underperformed on the ballot

One of the major techniques I’ve used for estimating the electability of various candidates is comparing them to other candidates on the same ballot.7 While Kamala Harris and Donald Trump were the only candidates listed on ballots in all fifty states, many other candidates ran for statewide office at the same time, including thirty-four running for federal office.

How to read the subsequent plots: Presidential candidates usually experience an overvote relative to downballot candidates (upper right quadrant). Evidence of undervoting and widespread ticket-splitting indicate a candidate is weak.

Hillary Clinton won and lost states in tandem with Democratic senators, although sometimes by considerably different margins. Joe Biden won one state that elected a Republican senator; Kamala Harris lost several key battleground states that nevertheless elected Democratic senators.

Clinton, Biden, and Harris versus senatorial candidates on the same ballot.

If we compare votes for senators and presidents - both federal offices! - Biden’s performance looks clearly superior to that of Kamala Harris. He earned more votes above the senatorial baseline, he held Trump to fewer votes above the senatorial baseline, and this is true whether we look at average or typical values. Typically, Biden performed 1.1% behind senatorial candidates, and Kamala Harris performed 2% behind senatorial candidates. If we focus on net percentage margins, Kamala Harris also underperformed Hillary Clinton.8

Now, there is a problem with this benchmark: These are three different collections of states. The Senate has a very curious rhythm to it related to incumbency and the countercyclic midterms. However, there are a group of high-profile politicians who face election in the same states at least once every four years: Governors.

Clinton, Biden, and Harris versus gubernatorial candidates on the same ballot.

There are some striking outliers, which gives a strong motive for looking at median values. Joe Biden’s median performance was 5.9% ahead of Democratic governors, while Kamala Harris’s median performance was 2.9% ahead of Democratic governors, a whopping 3% difference. Hillary Clinton’s median performance was 3.9% behind Democratic governors.

Now, that said, gubernatorial races can vary quite remarkably from normal partisan behavior. Kamala Harris’s worst battleground state performance relative to a gubernatorial candidate is a clear outlier: Josh Stein’s defeat of Mark Robinson in North Carolina was the most lopsided gubernatorial race in the state since 1980. However, North Carolina has a large number of statewide elections during presidential election years, including the entire Council of State.

Clinton, Biden, Harris, and Trump in comparison to 52 benchmark statewide candidates in North Carolina.

While Hillary Clinton showed signs of substantial overall undervoting, her typical net margin was 1% better than the typical Democrat running statewide in North Carolina, and the same was true for Joe Biden. However, Kamala Harris typically performed 1.5% worse - a gap of 2.5% between Harris and Biden relative to baseline partisanship in North Carolina.9

Democrats coming home: A deceptive polling surge

I’ve noted that polling showed many unusual characteristics this year. For example, polling overestimated Trump’s performance in presidential primaries while underestimating Biden’s performance. Kamala Harris’s approval rating shows a particularly interesting pattern:

Kamala Harris net approval, showing a sharp discontinuity circa July 21st, 2024, when Biden announced his withdrawal from the race.

At the beginning of the Biden administration, Harris’s net approval ratings generally lagged about 5% behind Biden. Over time, her approval ratings generally converged with Biden’s as Biden’s approval rating declined and reached an apparent partisan floor. Due to the fact that a significant but vocal minority of Democratic partisans wanted to substitute another candidate for Biden, I expected Democrats to “come home” once the nomination was settled, as has happened with both parties in previous intraparty conflicts.10

The moment Harris became the presumptive nominee, however, there was a sharp discontinuity in Harris’s approval ratings. It is likely that this discontinuity did not reflect any sudden change in approval of Harris among swing voters; instead, it probably reflected a shift in response rates that also impacted head-to-head polls.11

Candidate quality matters, not just platform

Blueprint 2024’s issue-oriented exit polling sought to identify the top reasons why voters voted against Harris: Cultural issues, inflation, and immigration. I’ll talk about this more in a later article but one key point I will emphasize is that “Democrats are too liberal” simply did not resonate with swing voters who cast ballots for Trump.12

The fact of the matter is that while Democrats as a party as a whole have embraced a number of unpopular positions on immigration and cultural issues and Democrats as the governing party have been blamed for inflation, these are problems that impacted downballot Democrats as well. I expected Democrats to stick with Biden as their nominee for three very concrete reasons:

On the fundamentals, replacing the presumptive nominee at the last possible minute is an extremely risky move.
If Biden were to voluntarily step aside, the only plausible choice was Kamala Harris, who was very clearly a weak candidate.
The Democratic Party has been very keen to avoid repeating the mistakes of 1968.

While we cannot know how Biden would have performed, it is now clear both that Democrats repeated several of the mistakes of the 1968 election and that Kamala Harris proved to be a weak candidate. Those who led the campaign to pressure Biden cannot reasonably claim to be vindicated in their decision.13

Not only has she become the first candidate to lose the popular vote to Trump, she did so in environments where in spite of any negative coattails she brought to the election, Democrats still were able to win statewide elections on the same ballot in key battleground states, including Michigan, Wisconsin, North Carolina, Nevada, and possibly Arizona.14 Candidate quality matters.

Thanks for reading MathIntee Political Post! This post is public so feel free to share it.

As I cautioned this summer. See “Echoes of 1892, 1948, 1968, and 1984” from June.

According to VoteView’s DW-Nominate, Kamala Harris’s score was less extreme than Elizabeth Warren and Sylvia Garcia.

As I discussed in “How to lose the 2024 election” in June, appealing to the median voter matters. On many issues where Trump and Biden differed with the median voter lying between the two of them, Kamala Harris put herself on the record in 2019 as endorsing an extreme policy. E.g., she pledged to decriminalize border crossings, rolling back a law that has been on the books since 1929; exit polling has highlighted that immigration remained one of the most important issues for swing voters.

"Electability and the Blue Bubble,” from October 2019.

Overall, she ended up with a proportionate volume of coverage, and the negative coverage near the end of her campaign may have made up for that. See "Media bias in the Democratic Primary,” November 2019.

"Electability and the Senate in 2020,” from January 2020.

Including both real elections and polls. Just as states have different levels of baseline partisanship, polls have different levels of baseline partisanship due to sampling methods and also random chance.

Hillary Clinton’s performance was more varied and a little harder to evaluate; she achieved the unusual feat of earning fewer raw votes than senatorial candidates, but while her net margins were high variance, they were actually perfectly in line with Democratic senatorial candidates. That said, Democratic

The gap between Biden and Harris increases to 3.0% if we use mean values instead of median values, but this is largely due to the aforementioned outlier. The Clinton-Harris gap stays constant at 2.5%, though, because there was a significant gap between a minority of higher-performing Democrats and a majority of lower-performing Democrats in 2016.

See most particularly Trump’s numbers in 2016.

This could help explain why polls consistently overestimated the protest vote against Biden during the primaries while simultaneously overestimating Kamala Harris’s support. An additional factor is that relative to Joe Biden, Kamala Harris’s popularity was higher among elites while lower among the working class, something exhibited to great effect during the 2020 primary cycle.

“Democrats are too liberal,” -4 among swing voters overall, vs +1 issue importance among voters overall. By contrast, “Kamala Harris is too liberal” registered at +4 with this group.

As a matter of historical reality, many of the key figures involved in the pressure campaign either publicly supported Biden’s nomination leading into the primary process or opposed Biden’s nomination in 2020 in the first place.

Arizona has not been called by the AP as of the time of publication of this articles. Given the assumption of coattails, a poorly-run presidential campaign could be blamed for losses in Pennsylvania and Ohio.

Back to the basics: Election forecasts

Tomas McIntee — Fri, 01 Nov 2024 01:53:33 GMT

As I have previously noted, there are numerous alarming signs that the quality of political polling has declined substantially. For the last several months, it has been my contention that poll-based forecasts of the election are likely to be lower quality than historical forecasts of the election based on the previous two elections.

That is not to say that presidential elections can usually be predicted well by the previous two elections. In fact, historical models of elections are generally bad, as the parties periodically realign themselves substantially. Even given that Trump was on the previous two Republican tickets, I do not expect the historical model to perform exceptionally well; instead, I expect that polling-based models have limited accuracy.

I’m putting this post out before the election to preregister my assessment of the quality of current polling. Because the main wild card in this election is the fact that Joe Biden stepped away from the nomination, the historical model also serves as a benchmark for the Democratic decision to run Kamala Harris.

The simplest models: Averages and trends

The simplest way to predict the result of any event is to guess that it will be roughly the same as the previous time. The next simplest way to predict events is by guessing that it follows a trend that changes over time, and forecast the trend based on two or more of the previous instances of that event. Usually, the best guess is somewhere between the trendline prediction and a simple average, but usually any prediction falling between an average and a linear trend is somewhat plausible a priori.

What I will do in this basic historical model is predict the results of the current presidential election based on the previous two presidential elections. This is necessarily a very simple model, with two steps: First, project the current vote for a party based on the previous election. Second, adjust the estimate for overall turnout.1

With only two previous points used for prediction, both trend and average models are just weighted sums of both previous points. For a trend model, this is a weight of 150% of the previous cycle and -50% of the cycle before that; for an average model, this is a weight of 50% for each of the two cycles.

Green density curve represents the trend model, which is worse than the average model. Errors exceeding 10% are not especially unusual

In objective terms, both of these models are terrible at predicting historical election results, even after throwing out all elections with more than two major candidates, the Reconstruction era, and the largely unprecedented 1932 election. Out of n=563 statewide elections for president in the 12 most ordinary-looking presidential elections, fewer than a third have results that fall between the predictions of the average model and the trend model.

This is, I will reiterate, a genuinely bad model. Even after focusing on the 12 most ordinary presidential elections in American history, the root mean square error of the trend model is 9.2% of the electorate, while the root mean square error of the average model is 7.8% of the electorate. A weighted average (75%-25%) has slightly better performance, with an error of 7.3% of the electorate.2

Predicting 2024: Illustrating the models

While this isn’t true for every state, the fact that Joe Biden performed considerably better in 2020 than Hillary Clinton did in 2016 means that the trend model predicts better Democratic results, while the average model predicts worse Democratic results.3 Between these is a natural parameterized range of linear historical models.

Reversion to the mean predicts an extremely narrow electoral victory for Donald Trump with 271 electoral votes and a minority of the vote; the trend model predicts a 303 electoral vote Kamala Harris victory similar to Biden’s 2020 map with a stronger popular vote margin.4 The most realistic weighted average model predicts a 276 electoral vote victory for Kamala Harris.

Judging the prediction models

For the purpose of calculating error in a prediction model, I will look at the root mean square error of the model’s central predictions of the popular vote in each state. My baseline prediction is that at least one model in my specified range of historical models will outperform some poll-based models.

Based only on the historical validation, I have selected the 75-25 weighted average model as most likely. If this model, specifically, outperforms most poll-based models, I can reasonably claim I’ve predicted the election better than the polls.

Thanks for reading MathIntee Political Post! This post is public and will embarrass me if you share it before I am proven wrong, so share away!

If none of the specified range of historical models outperform any of the poll-based models, then I clearly owe pollsters a collective apology for managing to overcome the extremely difficult polling environment with modern response rates.

Judging the choice of candidates

If Kamala Harris performs better than the historical range of models in a state, this would be evidence in favor of the switch having been a true game-changer. Similarly, Kamala Harris performing worse than the historical range of models in a state would be evidence that she was a poor choice of standard-bearer.

Not entirely independently, Trump’s performance can also be evaluated by the same benchmark. In 2020, Trump performed between the average and trend of the 2012-2016 cycles in twenty-six states, higher than the historical average of around one third of all states.5 My personal expectation is that as a third-time candidate, his performance will likely be between the average and trend of the 2016-2020 cycles in more than twenty-six states this time.

Whether or not the polls prove incorrect, I will follow up after the election by using the historical “previous two” election model as a method of estimating the strength of 2024’s presidential candidates as well as others over the years… so stay tuned.

This last step is necessary because elections sometimes involve large changes in absolute turnout numbers, in many cases due to rapid changes in the size of the eligible electorate.

By comparison, most polls have an expected error of less than 5% based on sample size… if one assumes that the respondents are a truly random sample of the population.

Donald Trump’s performance in both elections was approximately the same, although trends both up and down were visible at the state level.

Also, the trend model predicts negative votes for Donald Trump in Delaware, so it’s guaranteed to miss at least one state reasonably badly.

Interestingly, he overperformed in eighteen states and underperformed in six. The fact that this was not good enough highlights both how lucky Trump got in managing to win the Electoral College in 2016 given his overall level of popular support and how much more effective Democrats were in 2020 than in 2016 - the Biden campaign brought 2016’s third-party voters into the Democratic fold, at least temporarily.

Can we trust polls anymore?

Tomas McIntee — Sat, 17 Aug 2024 19:33:14 GMT

Previously, I’ve mentioned that this election cycle reminds me of 1948 in a number of ways, with Biden in the role of Truman. Now that Biden has decided to drop out, the parallels between 1948 and 2024 have become a little less salient to the public, but I still want to explore them - and, in particular, explore public opinion polling.

The polling associated with the current election cycle displays several warning signs that undermine their credibility as predictive instruments:

Unusual demographic cross-section results (weird crosstabs)
High levels of undecided voters
A consistent pattern of missing in the same direction over and over again during the primary season

But first, let’s step in our time machine, and go back to 1948 for some benchmark comparisons.

The 1948 election

One of the most famous headlines of all time comes from the 1948 election: DEWEY DEFEATS TRUMAN. Like Joe Biden, Harry Truman was less popular than other members of his party among the “chattering class,” with a number of energetic enemies in the press. Notably, this included the editor of the Chicago Daily Tribune, who confidently okayed going to press with a headline declaring that his detested enemy had been deposed before returns came in.

After all, polls showed that Dewey had a firm lead in popularity over Truman. Based on polling, Truman had a net disapproval rating of somewhere around 8%, and head-to-head polls showed him losing to Dewey by a similar margin. Some polls even showed him trailing by double digits.

Truman faced serious challenges from within his party by progressives (with FDR’s previous vice president, Henry Wallace, running under the Progressive banner) and segregationists (with Governor Strom Thurmond of South Carolina running under the Dixiecrat banner).

What happened with polls in 1948?

There are many accounts of why polls missed in 1948. The main retrospective report on the failures of polling in 1948 places the blame on a variety of factors. There were three main pollsters in the field surveying the public before the 1948 election: Roper, Crossley, and Gallup. All three pollsters used quota sampling, which was strongly biased in ways that led them to underestimate Truman’s support and overestimate Dewey’s support.

Additionally, Roper and Crossley did not run polls close to the election, and therefore missed by larger margins than Gallup due to late shifts of undecided voters. A large number of undecided voters were “double haters” who had previously voted for FDR, disliked Truman, but disliked Republicans even more.1

The 538 polling average at launch. Note that Trump’s average was five points below his 2020 performance; 7.5% of voters being “undecided” was fairly high, and RFK’s average of 10% was a priori likely to collapse.

2024 cycle polls face a similar set of problems. While potential respondents are generally selected randomly, response rates to surveys are extremely low, and this means that selection effects tend to dominate responses just as they did in 1948. Additionally, most polls showed an unusually high percentage of undecided “double hater” voters, as well as significant support for third-party candidates.

It’s plausible, given the decline in approval ratings, that Biden was actually on track to lose the popular vote to Trump by 1% with high rates of third-party voting; however, it is not plausible that 10% of the electorate lacked any preference between Biden (vice president 2009-2017 and president since 2021) and Trump (president 2017-2021). Yet 10% undecided was a pretty typical topline result for 2024 cycle polls!

The crosstabs were (and still are) crazy!

One thing that pollsters have some control over is trying to achieve a representative sample. In 1948, this was achieved by quota sampling. In 2024, this has been achieved either by the modern version of quota sampling or by reweighting samples in order to simulate a more representative population. The problem is, what we’ve seen from exit polls is that this has not been working very well. I’ve seen a steady stream of complaints about this on Twitter, because I follow enough polling geeks on Twitter.

Illustrative post from @forgetprotocol on Twitter.

From 2012-2022, pre-election polls have predicted a doubling in the Republican share of the African-American vote from 6% to 13%, and a decline in the Democratic share of the African-American vote from 93% to 75%. Exit polls, on the other hand, suggest that Democrats won the African-American vote by a margin of 93% to 5% in 2022.

Polls have shown similarly dramatic shifts in the age profile of support for Democrats and Republicans. A significant number of polls conducted in 2024 showed Biden winning voters over age 65 while losing voters age 18-29, a dramatic reversal from what exit polls showed in 2020.

What about Biden and Trump specifically?

We’ve actually had a significant set of opportunities to compare polling to outcomes for Biden and Trump separately, as both Republicans and Democrats have had a series of primary elections. There have also been polls attempting to capture this.

112 presidental primary polls, plotted by density of error.

In 538’s dataset, there are 112 primary polls conducted after January 23rd, 2024 and did not test hypothetical candidates.2 Polls targeting a specific state tended to be somewhat more accurate, but the overwhelming majority of polls underestimated Biden’s support and overestimated Trump’s margin over Haley. If we focus on “likely voter” polls and eliminate one particular consistently-wrong tracking poll to eliminate the visible spikes in the distribution, the pattern remains in the much more selective set of 58 polls:

In other words, at the same time that national polls indicated that Biden was behind Trump by 1-2%, the very same polling outfits were demonstrably reaching samples that were more in favor of Trump and less in favor of Haley and Biden. While it is possible to generate ad hoc explanations for the systematic errors, the simplest explanation is that polls are struggling to predict voting behavior accurately.

Why polls are cooked in summary

First, we’ve seen significant indications that polls are no longer capturing a representative population. Old respondents (65+), young respondents (18-29), and African-American respondents in particular have shown substantial and growing divergence between polling crosstabs and exit polling in the post-2012 era. Attempts to control by education and income have not prevented a skew that appears to be linked to socioeconomic status and rates of online activity.

Second, the projections derived directly from political polling presented implausible voter behavior. There is very little reason to believe that an election between Trump and Biden would result in a popular vote of 41.6% to 40.9%; pollsters needed to do a better job of resolving the breakdown of “double hater” voters’ behavior in order for their polls to support prediction.

Third, we concretely have seen a pattern of substantial errors in polls underestimated opposition to Trump from Republican voters and overestimating opposition Biden from Democratic voters. This seems to have gone largely unnoticed by media outlets that have generally provided disproportionate coverage to opposition to Biden among Democrats (e.g., Rashida Tlaib & pro-Palestinian protestors) over opposition to Trump among Republicans (e.g., former president George W. Bush, former vice presidents Dick Cheney and Mike Pence, and half a dozen sitting Republican senators).

This does not mean that public opinion polling is completely useless; however, we have good reason to believe that it is becoming increasingly unreliable when it comes to subjects where elite opinion diverges from popular opinion. Pollsters have tried to correct for the errors in partisan toplines that understated Trump’s support in 2016 and 2020; but the modern version of quota sampling is subject to strong selection effects every bit as strong as walking around the streets in 1948.

Southern Democrats in particular - Dewey was a New Yorker and fairly liberal.

538’s data does not record any projections for “uncommitted” that I am aware of.

Defining the Horseshoe Caucus

Tomas McIntee — Sat, 13 Jul 2024 18:17:49 GMT

Horseshoe from left tip to right tip: Bernie Sanders, Alexandria Ocasio-Cortez, John Fetterman, Kyrsten Sinema, Joe Manchin, Lisa Murkowski, Mitt Romney, Marjorie Taylor Greene, and Rand Paul.

The most common way to organize political perspectives is along a one-dimensional spectrum. The traditional organization, dating from seating arrangements in the National Assembly of France in 1789, places those opposed to change (conservatives) on the right, and those in favor of change (progressives) on the left.

Subscribe now

Oddly, however, it’s not unusual for “far-right” and “far-left” figures to find common cause against the center; this has been happening in the United States Congress lately, leading to the rise of a visible group that I refer to regularly as the “Horseshoe Caucus.” This article will describe and define the caucus more formally.

Horseshoe theory

In general, American politics tends to be organized around a bipolar political spectrum, usually referred to as the “left” and “right” in a convention for political discourse that can be traced to seating arrangements adopted by French political parties during the Revolutionary era.1 Following the slow exit of conservative Southern Democrats from the scene, most mathematical analyses suggest that on the vast majority of issues, Congress is strongly polarized in a single salient dimension.

Many students of politics, however, have been puzzled by the occasional intersections of interest between the far left and far right. The observation that far-left and far-right figures tend to find common cause against moderates of both wings is sometimes called “horseshoe theory.”

Building the model of the horseshoe

In order to consistently and mathematically analyze Congressional votes to identify the Horseshoe Caucus, I will be utilizing Voteview roll call data and categorizing votes by type based on the results. Many votes are strictly party-line affairs, and tell us nothing about the Horseshoe Caucus.

Many votes are carried by a partisan majority with some defections. If the parties are not perfectly polarized, these defections may cross the aisle in both directions. I will use these votes in order to build a main “left-right” spectrum from Democratic to Republican.

There are also quite a few votes where a majority of both parties are aligned. These votes can be distinguished by considering the partisan composition of the minority: If the minority belongs to only one party, the vote still fits neatly in a one-dimensional political model. Votes with a bipartisan majority and bipartisan fringe opposition are “horseshoe” votes that identify members of the Horseshoe Caucus.

Is there a horseshoe?

The short answer is “yes.” Below is a plot of the frequency of “horseshoe” votes for the members of the 118th Congress during the period of the 115th-118th Congresses compared to alignment in normal partisan votes.2 The best LOESS fit to this is a U-shaped curve, although this has some problems.3

It can be a little difficult to see exactly what’s going on because partisan loyalty is highly skewed - most members of Congress are highly loyal to their party during normal partisan votes. Below, I’ve changed partisanship to an ordinal variable - that is, ranking the members of Congress by their partisan loyalty in normal partisan votes - and highlighted a few members of Congress in different regions of the graph.

Out of twenty-five members of Congress with horseshoe factors of exactly zero, there are only two with higher than average party loyalty: John Fetterman and Vince Fong.4

The Democratic members of the Horseshoe Caucus are all strong partisan loyalists. Alexandria Ocasio-Cortez crosses the aisle most often, but still scores an above-average 97% on Democratic partisanship along with nominally independent Bernie Sanders.

Republican members of the Horseshoe Caucus include some with low partisan loyalty, most notably Thomas Massie (87% Republican adherence in partisan votes while casting horseshoe votes against a bipartisan majority a staggering 65% of the time). Republican Vince Fong is notable as the only member of the 118th Congress with perfect partisan loyalty on normal partisan votes; he has also completely avoided casting any horseshoe votes.

What is the composition of the Horseshoe caucus?

Within the House of Representatives, most of the Horse Caucus are either Justice Democrats, current members of the Freedom Caucus, or former members of the Freedom Caucus. There are twelve Justice Democrats in total; the twelve House Democrats who cast horseshoe votes most frequently include eleven Justice Democrats and all nine members of The Squad.5

While the House Freedom Caucus does not publicize its member list, around 90% of House Horseshoe Caucus Republicans are known to be members or former members. Similarly-aligned Trumpist Senators make up most of the rest of the right wing of the Horseshoe Caucus.

Moving back to the other side of the aisle in the Senate, four of the five progressive senators who ran against Joe Biden for the Democratic nomination: Bernie Sanders, Elizabeth Warren, Cory Booker, and Kirsten Gillibrand. (The fifth, Kamala Harris, was a staunch member of the caucus during the 115th and 116th Congresses with a 56% horseshoe rating.) Among those five, Cory Booker scored the lowest in post-Trump horseshoe voting at a still-impressive 38%.6

Rounding out the rest of the Democratic side of the aisle in the Horseshoe Caucus are Elizabeth Warren’s fellow Massachusetts senator (Ed Markey) and both of Oregon’s senators (Ron Wyden and Jeff Merkley). The highest horseshoe ratings, however, belong to a pair of Kentucky Tea Party Republicans, Rand Paul in the Senate and Thomas Massie in the House.

Methodological caveats

It is worth noting that as a group, the Horseshoe Caucus remains overwhelmingly partisan, voting in line with their respective parties over 90% of the time. However, there is a small identifiable group of politicians who frequently vote against bipartisan majorities, and that group includes both left-wing and right-wing figures.

The methods I’ve used to identify the Horseshoe Caucus rely heavily on explicit partisan identity; a horseshoe vote is defined as having at least one member of both parties in the dissenting minority. This is mostly non-problematic in the modern era of very high polarization.7 As a robustness check, I’ve run the data back as far as 1994; the presence of RINOs and DINOs becomes visible on the partisanship axis in the 1990s, although the distribution of scores still suggests a horseshoe shape.

A common criticism offered in analyzing politics in a two-dimensional or higher-dimensional spectrum is that the dimensions may not be independent. However, in this case, the 3,637 votes I used for analyzing partisan loyalty do not include any of the 500 votes I used for measuring horseshoe voting against a bipartisan consensus.

These measures are mathematically independent; if the most visible other effect was simply the degree to which members obeyed party discipline, we should see a chevron shape anchored by politicians similar to Vince Fong, who vote with the majority of the party in both types of votes, with bipartisan dissents being most commonly found among bipartisans like Joe Manchin and Lisa Murkowski.

Summary

If we look at voting patterns within the 115th-118th Congresses, there are a significant number of members of Congress who join in bipartisan opposition to bipartisan majorities. These members are disproportionately located on the hard left and hard right, and are firm partisan loyalists when it comes to the standard partisan issues… which shows a horseshoe.

I will leave an in-depth exploration of the issues that unite Justice Democrats and the Freedom Caucus for later; but I will note one area of interest is foreign policy. Many of the usual suspects voted against the “TikTok ban” law and have voted against aid for Ukraine, Israel, and Taiwan.

While during the French Revolution, “liberal” was synonymous with “left,” subsequent evolutions in the political spectrum mean that “liberal” tends to mean something quite different from “leftist.”

Specifically, partisan votes with some defections. Note that the 115th-118th Congresses correspond to the Trump and Biden administrations, with Trump-related realignments 2016-2018 being visible in the data.

Note in particular the negative values in the middle.

The full list includes eight other Democrats, fourteen other Republicans, and independent Angus King (ME-Sen).

The exception within the Justice Democrats is Ro Khanna, whose horseshoe score is a mere 18% - still high, but not exceptional. The exception within the Horseshoe Caucus is Chuy Garcia, who joins bipartisan dissents against bipartisan majorities 28% of the time, putting him in eleventh place among House Democrats.

Within the pool of Trump-era votes, Kamala Harris had a horseshoe score of 56%.

Arguably, it is too restrictive.

Echoes of 1892, 1948, 1968, and 1984...

Tomas McIntee — Thu, 04 Jul 2024 18:14:59 GMT

Remixed from source. Other images in this article are public domain or own work.

My book on the Electoral College discusses all fifty-nine presidential elections from 1788 to 2020 because I wanted to have a broad and historically-grounded perspective on how the Electoral College works. Today, I want to talk about which presidential elections I’ve been thinking about as the 2024 election draws close.

Naturally, this includes recent elections - Trump and Biden are familiar figures in American politics - but also the elections of 1892, 1948, 1968, and now 1984. Each of these elections has key lessons to teach us about this election and this moment, both in terms of political strategy and in terms of expectations.

From a historical perspective, I think Biden’s chances of victory against Trump are widely underrated; relatedly, the importance of anti-Biden dissent is overrated while the importance of anti-Trump dissent is overrated; and the consequences of abandoning Biden or Trump as nominee are vastly underrated.

The big picture

Right now, the big picture is that both Trump and Biden are polling consistently under 50% of the electorate. Neither candidate is particularly popular right now, which is a situation similar to 2016.

Rematch elections tend to have similar results, and this should make the election relatively easy to predict.

Echoes of 1892

Only one former president has ever returned from office: Grover Cleveland. This happened in an interesting way:

1884: Grover Cleveland won a plurality of 48.8% of the vote.
1888: Grover Cleveland won a plurality of 48.6% of the vote.
1892: Grover Cleveland won a plurality of 46.0% of the vote.

Grover Cleveland won the popular vote three times. Prior to his first presidential run, he also won the popular vote by an exceptional margin when running for governor of New York:

Donald Trump is not Grover Cleveland. However, like Grover Cleveland, his support has been relatively stable thus far: He lost the popular vote twice with 46.1% and then 46.8% of the vote. His current polling average is actually below 46%, with a large undecided share of the population; it is quite likely that he will earn less than 47% of the vote in 2024, though it remains possible for him to win with as little as 46% of the vote. As in 1892, winning with 46% of the vote requires that a significant share of voters choose to reject both parties.

Candidates rarely go from losing the popular vote to winning it. William Jennings Bryan, another three-time candidate, went from 46.7% of the vote to 45.5% to 43.1%. His first two contests were matches against the same opponent (William McKinley). The history of rematches suggests that Donald Trump’s level of support in 2024 should be very similar to his level of support in 2016 and 2020.

The big question is this: What about Joe Biden? How much of his 2020 support can he retain? Will voters choose to support him against Trump - or will they abandon him, voting for third-party candidates, leaving the top line of their ballot empty, or holding their nose to vote for a candidate they detest?

1948 and the Dixiecrat rebellion

In 1948, segregationists from the Southern wing of the Democratic Party stormed out of the Democratic National Convention and nominated their own presidential candidate, Strom Thurmond. At this point in time, the segregationists of the Solid South were not swing voters in the two-party system. They were politically powerful, but represented the views of a small minority of voters whose perspective was (at the time) completely excluded from the Republican Party.

Biden has also faced loud and extremely passionate criticism from a small segment of the American population that rallied in support of Gaza following Hamas’s attacks on October 7th, 2023 and dislikes Biden’s moderate position on Israel. Like the passionate segregationists, this segment makes up a minority of the Democratic Party and is not welcome in the Republican Party; and supporting Donald Trump on behalf of Palestinian interests makes as much sense as supporting Thomas E. Dewey on behalf of segregationists.1

While under fire from segregationists, Harry Truman’s approval rating fell to under 40%, a figure that seems all the more impressive when we put it in the context of the fact that his peak approval rating had been north of 80%. Polling generally showed Truman substantially behind, sometimes thought to be an artifact of socioeconomic factors and uneven adoption of new technologies.2 Compounding this factor, journalists and the educated elites that news editors generally rub elbows with disliked Truman, mostly for reasons related to class, identity, and culture rather than partisan affiliation.

Today, as in 1948, a substantial share of partisans whose natural home remains in the Democratic Party overtly and loudly disparage Biden, demanding that he abandon principles that enjoy bipartisan popularity, and pollsters struggling with uneven adoption of new communications technologies find that they are having trouble reaching representative samples. Meanwhile, journalists and the educated elites that news editors generally rub elbows with dislike Biden, mostly for reasons related to class, identity, and culture rather than partisan affiliation.

Sound familiar to you? It sounds familiar to me.

The Democratic National Convention in Chicago

This year, Democrats are holding their national convention in Chicago. One of the most famous Democratic National Conventions happened in Chicago in 1968. It wasn’t pretty, and it wasn’t pretty for an assortment of reasons: Incumbent Lyndon B. Johnson stepped away from running again, one of the arguable frontrunners (Robert F. Kennedy) was assassinated, and then the delegates at the convention decided to nominate Hubert Humphrey, who hadn’t participated in any of the state primary contests.

Old conservative Southern Democrats were angry because Hubert Humphrey was the firebrand anti-segregationist who set the 1948 convention on fire. Young liberal Democrats were angry because Hubert Humphrey supported the Vietnam War. There were protests. There were riots. Segregationists ran their most successful third-party bid yet; with a geographically diffuse base of support, Hubert Humphrey lost decisively in the Electoral College, although the actual popular vote was quite close.3

Today, two separate groups seem intent on trying to make Chicago 2024 a repeat of Chicago 1968: Leftist protestors who want to force the Democratic Party to pivot hard to the left (the war in Gaza being one of the loudly focal subjects) and elites who want to force the Democratic Party to abandon their elected nominee in favor of holding the first brokered convention since 1968.

The 1984 debate

The first presidential debate was held last week. Both candidates showed visible signs of aging poorly, offering answers that sometimes veered into confusion and incoherence and sometimes were completely unrelated to the topic at hand. However, they were understandable most of the time, and reading the transcript shows little difference in levels of incoherence between the two. (Full debate transcript available here.)

Biden reportedly had a cold; he was quieter and visibly less energetic, particularly during the first half of the debate. (During the second half, Biden became more energetic and Trump became less energetic, narrowing the gap in energy levels substantially.)

In 1984, an aging Reagan showed signs of age, forgetfulness, and incoherence in the debate. Some scholars have argued that these were the early signs of Alzheimer’s on display; some have said this was normal age-related decline in performance. After the first debate, pundits were shocked, and declared Reagan’s campaign dead in the water… and then Reagan went on to win the most lopsided electoral victory of the last half-century.

Concerns over Reagan’s age were legitimate, as are concerns today over the age of Trump and Biden, both of whom are older than Reagan was in 1984. However, Biden having a poor debate night during which he showed more signs of age than usual is not likely to have (by itself) any major impact on voters’ choices. Instead, what is important are the secondary effects; if Democrats force Biden out, it is likely to have disastrous consequences.

Recap and review

At present, Donald Trump has led in polling for most of the year, but it is still several months before the election, and polling indicates that there are a large number of voters who are not yet ready to say that they will support either candidate.

There are good fundamental reasons to expect that Donald Trump will struggle to win more than 46% of the vote; there are also good reasons to believe that, between the dislike he has attracted from the pro-Palestine left, journalists, and elites, media coverage (and, for that matter, polling), it is easy to underestimate Biden’s support at the end of the day in the general election.

While both major party nominees are, frankly, too old to make it advisable to elect them to the position of President of the United States of America, neither party is in a position where they are likely to win by throwing out the results of their own presidential primaries. The balance of the election rests in the hands of swing voters who are not well understood by the reliable partisans who dominate both “blue” and “red” flavors of media.

Dewey was a New York Republican. His platform included support for a federal anti-lynching law, abolition of poll taxes, opposition to racial segregation in the armed forces, and civil rights legislation to support equal opportunity regardless of race. His ticket and platform was total anathema to segregationists, who found him no more palatable than Truman.

His running mate, Californian Earl Warren, later became the chief justice of the Supreme Court from 1953 to 1969; he authored the Brown v. Board decision in 1954 and was instrumental in numerous civil rights cases. Another repeat loser of the popular vote, Dewey earned 45.9% of the vote in 1944 and 45.1% of the vote in 1948.

Household adoption of telephones was uneven, so telephone polling skewed heavily to higher-class households, which skewed poll results in favor of Dewey.

The popular vote margin was under 1%.

How to lose the 2024 election

Tomas McIntee — Mon, 01 Jul 2024 00:53:21 GMT

Speaking as an expert, I will say this: It is too early to confidently predict the results of the 2024 election. However, in line with what I have said in the past about the median voter and electability, what we do have a good idea about is the shape of public opinion, the positions of the candidates, and where those candidates are vulnerable. Today, I’m going to tell both parties how to lose the election.

This could be a close election either way; it could be a landslide defeat. The key to losing the election is alienating and offending as many voters as possible.

Double down on controversial victories

True popular progress on social issues comes slowly in either direction; both parties have adopted numerous hardline positions that are not popular with moderate voters. This creates a “pendulum swing” effect: Each side undermines their own popular support when they score major victories.

Republicans recently won a major victory on abortion with Dobbs v. Jackson Women’s Health Organization (2022), which repealed Roe v. Wade (1973). Since the median voter supports moderate Roe-compliant restrictions on abortion, Republicans can alienate numerous swing voters by calling for a total national abortion ban.

While Democratic movement away from the winning position of “safe, legal, and rare” almost certainly undermined Democratic support in the 2012-2020 cycles, it’s less clear that Democrats have an easy path to alienating voters nationally on the abortion issue in a post-Dobbs environment.1

For Democrats, the equivalent of abortion issues are trans issues. Democrats have scored major victories in gaining legal and institutional support; however, transwomen competing in women’s sports polls at a net negative 41 points2 and a majority of the population thinks that changing gender is morally wrong.3

Support Russia, China, Iran, or Hamas

At present, polling shows net negative favorability rating for Donald Trump (-12.1%) and Joe Biden (-15.6%). On the balance, this implies that Trump is favored by a net share of about 1.8% of the population. On the same scale, the gap between favorable sentiments of Americans toward foreign countries in conflict is much more dramatic:4

Ukraine is favored over Russia by a net 61% of the population.
Taiwan is favored over China by a net 55% of the population.
Israel is favored over Palestine by a net 50% of the population.5

Both parties include fringe elements that are friendly to Russia, China, Iran, and Hamas, represented in Congress by what I like to refer to as the “Horseshoe Caucus.” The Horseshoe Caucus frequently bands together against bipartisan consensus positions, such as banning TikTok.6 On foreign policy, the Horseshoe Caucus presents different liabilities for different parties.

A majority of Americans support all three of Ukraine, Taiwan, and Israel. A minority of Americans support any of Hamas, China, or Russia.

The Democratic side of the Horseshoe Caucus - spearheaded by Rashida Tlaib, who is actively campaigning against Biden - is ready to undermine support for Democrats at the ballot box with overt sympathy for Hamas. This risk is amplified by key unelected Democratic elites, including within the “blue bubble” media ecosystem and on college campuses, where overtly pro-Hamas protests can be easily linked to the left wing of the Democratic Party.

The Republican side of the Horseshoe Caucus has been openly sympathetic to Russia and in favor of cutting aid to Ukraine. Donald Trump was impeached in 2019 for withholding military aid to Ukraine. The “red bubble” media ecosystem includes overtly pro-Russian figures like Tucker Carlson.

While both parties can lose the election by bending to the will of their side of the horseshoe caucus, Democrats have additional liability: As president, Biden will be held largely responsible by voters for the course of the wars in Gaza and Ukraine. No matter what Biden says, many voters will believe him responsible if the wars are going badly for Ukraine or Israel; thus, Democrats seeking to lose the election are well-advised to delay aid to Ukraine and Israel, holding it up with procedural measures or demanding restrictions on its use in the field.

Abandon the elected nominee

In theory, both parties have a chance to change out their designated candidates at the last minute at their party conventions.7 However, the voters have had their say, and any overt attempt to force out the elected nominees would be disastrous for the party that tried it.8

While Trump was clearly a poor choice of nominee, that choice has been made. Similarly, while “blue” media elites have revolted against Biden’s candidacy repeatedly, Democratic voters have stood behind Biden even more firmly than Republican voters stood behind Trump.

Remixed from source.

The nominees are locked and loaded. However, the elections of 2024 do not only include a presidential election. Both parties are contending for control of both houses of Congress, eleven state governorships, and 85 state legislative chambers in 44 different states. The last time a party threw its own primary results out of the window in favor of a candidate who did not win the primaries (Democrats in 1968), the party in question lost five senators and five governors in a downballot bloodbath.

Attempting to unseat the party’s nominee and actively campaigning against said nominee is generally an effective method of losing the election.

Embrace ideological purity, reject the median voter

Most of the advice I have given above boils down to this: Embrace extreme positions. Ignore what the voters say they want; instead, listen to elite opinion-mongers with large media platforms who reach one or another specific ideological audience.

The key to losing elections is to ignore swing voters, who lie within a fairly quiet and non-ideological segment of the electorate. Swing voters like to see stable institutions. They like political parties to respect the will of the voters participating in primary elections, even if they don’t much participate in those primaries themselves. And they fail almost every ideology’s purity tests.

If you are interested in losing, therefore, you should pay attention to the loudest voices on Twitter, the NYT op-ed page, Fox News, and NPR, and not so much to the boring nitty-gritty of public opinion polls; and to political scientists who shout about embracing radical positions to generate enthusiasm rather than those who tell say the median voter needs to be persuaded.

2012 is when the Democratic Party expunged its “safe, legal, and rare” plank; this also marks the pivot point on many other issues.

Pew poll here. This also shows a trend against basic acceptance that changing one’s sex is possible at all and more people who think that attitudes are changing too quickly than are changing too slowly.

See Gallup here.

Figures here are from Gallup’s 2023 polling. Note that the Palestinian Authority is the more sympathetic of the two Palestinian governments, with most indications being that American sympathies have shifted further toward Israel in 2024.

50% is using Gallup’s figure for the Palestinian Authority, which is fading from public view with the focus on Hamas in Gaza. This poll, focused on the war itself gives a narrower gap of 32% for validity of the conflict and 33% on the tactics used - in spite of the fact that most mainstream news coverage in the US comes from the “blue bubble” with a visible anti-Israel bias.

I don’t have the space in this article to spare for my full exploration of the Horseshoe Caucus. The underlying voting patterns are very interesting, though.

Note that ballot access laws and party rules make this move at best fraught with difficulty.

N.b.: I wrote this before the first presidential debate fueled a fresh round of demands for Biden to drop out from the NYT set. It’s still true as I hit the “publish” button today.