The 2028 campaign has already started. AOC is freezing her eggs, pumping iron, and winning Congressional allies. Highly engaged DSA members are planning varied endorsement processes, which will likely culminate in an endorsement of AOC’s presidential campaign soon.

Overlaying this background, DSA has released a platform, which is the product of a long democratic process within the organization. It is a mix of short- and long-term advocacies.

Tension between the stated, emergent agenda and the manichean imperatives of campaign silly season is inevitable in the AOC endorsement process and beyond. But the binding or advisory nature of the agenda currently operates as a proxy battle about DSA’s relation to the upcoming 2028 presidential primaries.

Without recounting specifics, my sense is that those who are most skeptical of endorsing AOC are the most sympathetic to the argument that AOC should adhere to our platform if she is to be endorsed. And those who are most enthusiastic about endorsing AOC are the least sympathetic to the argument that adherence to the letter of our platform should be the criterion for endorsement. But this tension over the so-called “red lines” hides a point of consensus: No major voices in DSA dispute that DSA will not have the power to control or discipline a top tier presidential candidate in the 2028 cycle.

Would be nice, but will not happen.

I propose, therefore, that DSA caucuses should strive not to have a lively debate about the proxy issue of the released platform. It factually won’t be imposed. If the side that wants it imposed wins on those terms, the endorsement will not happen (because AOC probably will not seek endorsement). If the side that wants endorsement beats imposing the platform as a condition, the endorsement will proceed with no firm commitments from AOC. Either way, the platform remains advisory.

There is a more constructive possibility.

It is better to think in terms of what, if anything, DSA’s active and broader membership wants injected into the Democratic primary process that is uniquely a socialist issue—something that isn’t the focus of Never Trumpers, or liberals, or even the best progressives. An issue that would attract broad support now, but that is also ripe for imagining a much brighter future won by the workers’ own hands. An issue that would both increase AOC’s attractiveness as a candidate and would also place more pressure on her to deliver once she wins. An issue the national media will not cover without the heft of a major organizing push directed at the heart of the attention-grabbing presidential race.

The issue is a labor agenda, specifically reversing the decline in union density, which the 2020 Bernie campaign promoted as a plan for Workplace Democracy.1

DSA has been involved in two types of presidential primary efforts in recent years: DSA for Bernie and Uncommitted. Although many consider the failure to field a 2024 candidate a major DSA blunder, one thing it did teach us was how to get millions of votes for an issue *alone*. Combining the candidate advocacy of DSA for Bernie and the issue advocacy of Uncommitted is a constructive possibility that will bring out the best in all of DSA’s current members and capacities.

Popularizing a labor agenda cannot be an afterthought. Every salt will be needed. Every union rep and every union staffer in the organization will have a role. Every Marxist and every street protest leader will be needed to think through how we can really supercharge this issue in the public imagination, and every one of them agrees that workers in motion, with or without AOC as president, is the most important short term goal for the socialist movement. Every veteran electoral campaigner will have a role designing townhalls and barnstorms where workers will be invited in to hear about AOC’s labor plans and the union density agenda.

These workers will turn into electoral margins in key primaries, and the centrality or contribution of the Workplace Democracy campaign to AOC’s campaign will come into focus. The focus cannot occur without people in motion around this issue.

DSA could endorse AOC along with a directive to NPC to run the AOC campaign on the basis of the Workplace Democracy agenda. If the legal structure we pick allows coordination, the directive should direct the NPC to directly coordinate with AOC guaranteeing DSA a position in the planning of her public campaign on this specific agenda item.

If DSA is not in a place to discipline, but also does not want to simply back a populist campaign with no tie-in to socialism, we should seek out this type of constructive engagement.

Endorsing with an offer—and commitment—to run an issue campaign with a crucial reform in addition to support for AOC would help to get DSA on the fighting, visionary footing we need both to actually get AOC through the primary and elect her president.

While I disagree with many aspects of what AOC’s left critics often say about her, there is this issue area the entire organization has never laser-focused on. What is going to be the relationship between a progressive/social democratic president and the effort to increase union density in the United States? DSA has an opportunity to leave it all on the field and shape the answer to that question even if we cannot grab control of a top presidential contender.

I recently read the book written by Daniel Tutt and C. Derick Varn2 cautioning against a full embrace of left populism. They spend a considerable part of the book explaining the deficiency of left populism—so far—to improve the union density picture. One cannot argue with the data, but then again there is no reason why the DSA portion of the broader push for AOC cannot have a decidedly socialist focus by placing union density plans on the national agenda.

AOC for President and Workplace Democracy for All!

More recently, the left has also attempted to learn from the right. Liberal activists created the State Innovation Exchange (SIX) as a left-leaning version of the conservative American Legislative Exchange Council (ALEC). Farther left, socialists have argued recently that the left should copy the parliamentary tactics of the congressional far right. Much of this discussion has focused on the House Freedom Caucus (HFC), the right wing caucus founded in 2015.

All the way back in 2018, Alexandria Ocasio-Cortez herself contrasted the unity of the HFC with the much bigger but more loosely organized Congressional Progressive Caucus. In 2020-2021, the “force the vote” debate on the left centered around whether socialists in congress should, like the HFC, threaten to vote against their party’s Speaker candidate (in this case Nancy Pelosi) unless they received a promise that one of their key legislative priorities (Medicare for All) would be brought to the floor for a vote. More recently Neal Meyer1 has argued that the HFC’s success in bullying Kevin McCarthy provides a model for the kind of combativity and cohesion that we should expect from socialist legislators.

I’m not convinced.

In my view, the HFC has been considerably less successful than is generally appreciated, and the costs of their strategy have largely been ignored. Their example for socialist legislators is, therefore, largely (but not entirely) a negative one.

Before getting into the comparison, it’s worth establishing what exactly the House Freedom Caucus is. The caucus was founded in January 2015. The impetus for its founding was the failure of far right congressmember Mick Mulvaney to win leadership of the Republican Study Committee (RSC). Founded by new right impresario Paul Weyrich in the 1970s to pull the party right, by the 1990s the RSC had become a central Republican leadership institution in the House. During the Obama administration, frustration with the RSC had mounted among the most conservative House Republicans, as the committee tended to support speaker John Boehner in cutting budgetary compromises with the Obama administration rather than shutting the government down. When Mulvaney lost the RSC leadership contest to Texas energy executive Bill Flores, he and his ideological confreres determined that a new organization was needed for conservatives unwilling to compromise with liberalism.

While the RSC leadership vote was the proximal cause, the founders of the HFC were also reacting against the failure of the House Tea Party Caucus. This caucus had been founded in July 2010 as the organization of the new belligerent right that was then storming Republican primaries. But despite the attention the caucus and its members commanded, by summer 2012 the organization was defunct, having failed, as an organization, to have much influence on anything.

Mulvaney and his comrades, notably Jim Jordan of Ohio and Raul Labrador of Idaho, were determined to avoid a repeat of the Tea Party Caucus. Instead of a group that was publicity seeking and undisciplined, they wanted a group that would be secretive and highly coordinated. To that end, they shut out House members like Louie Gohmert and Thomas Massie, who, while deeply conservative, had reputations as undisciplined loose cannons. The new caucus would be invitation-only and, unlike most caucuses, would not make its membership public. It would also bind its members by rules like the 80% threshold, which held that if 80% of the caucus supported a vote, the remaining 20% would have to follow.

Very quickly, the caucus became a thorn in Speaker John Boehner’s side, denouncing him and voting against bills like DHS funding and trade deal authority for the president. Boehner used his power as Speaker to retaliate, removing HFC members from subcommittees and cutting some off from National Republican Congressional Committee funds. In July, Mark Meadows filed a motion to vacate the chair, calling for a vote on Boehner’s speakership in the middle of the term. It was an extraordinary move, the first such motion in nearly a century. In order to avoid a bruising leadership fight that might end in defeat anyways, Boehner announced his retirement. He was replaced by Paul Ryan.

In the 2016 presidential race, the HFC didn’t play much of a role. The group was opposed to Trump, who they correctly saw as totally uncommitted to fiscal conservatism. They were largely divided between Ted Cruz, who had built a strong connection with the HFC, and Rand Paul, whose dogmatic libertarianism appealed to several of the group’s leaders. Though discipline and unity had been the caucus’s founding watchwords, it did not attempt to make a coordinated intervention into the Republican primary.

When Donald Trump became president, the posture of the group shifted considerably. While they continued pushing for far right priorities, they did not engage in the kind of all out warfare they had directed against Boehner. Instead, the group often compromised, and when they did fight, they often lost.

On taxes, perhaps the cause closest to the caucus’s heart, there were two priorities in 2017. First, they wanted Trump’s tax cut to be deficit neutral so it wouldn’t increase the deficit. In this, they were in accordance with the House leadership. Paul Ryan wanted the same thing. But they were opposed to Ryan’s means of paying for it - a border adjustment tax (BAT) that would amount to a huge tax increase on importers. This formed their second priority. Ryan would junk the BAT in summer 2017, though the HFC could hardly claim sole credit. The tax spurred opposition from a huge sector of business, from retailers to the Koch brothers, the latter of which used their connections with the Trump White House to turn the administration against the BAT as well. But with the BAT gone, there was no way to make the tax cut deficit neutral without huge spending cuts that would derail the entire bill. The HFC pushed for these, but got nowhere, and the Tax Cuts and Jobs Act passed without any compensating spending cuts or new revenue.

The second major fight for the HFC was Obamacare repeal. The caucus wanted a full repeal. Here, the HFC did come into serious conflict with the Republican leadership and the Trump administration, and were regularly denounced by both. Two members actually left the caucus in the midst of this battle. Paul Ryan had developed a bill, the American Health Care Act (AHCA), that would repeal much of Obamacare, but left in place some of the original bill’s new taxes, and also offered refundable tax credits to help households purchase health insurance. The HFC refused to support the bill, and lacking the votes to pass it, the house leadership canceled their scheduled vote on it. They subsequently passed amendments that would bring the bill closer in line with what the HFC wanted. Though they didn’t win all of their demands, the HFC agreed to support the bill. Mark Meadows explained that “had we not come around ... there would have been the typical stereotype that they’ll never get to ‘yes.’” But even in compromising, the HFC was unable to achieve its demands. The revisions to the bill had made it too conservative for moderate Republicans, who pulled their support, dooming the AHCA once and for all.

When the Democrats took back the House in 2018, the HFC faded in prominence. The reason is simple. To be a minority of a majority is to have some power. But to be a minority of a minority is simply to be powerless. From 2019-2023, the HFC voted basically the same as other Republicans. It joined the fight to defend Donald Trump against impeachment, causing one of its members, Justin Amash, to leave, and also worked to punish Liz Cheney for voting for impeachment after January 6. But neither of these distinguished the HFC from the bulk of house Republicans. When their party was in the minority, the caucus simply did not have very much power.

All of this changed after the GOP took the house back in 2022. As soon as the new congress came in, the HFC was determined to exercise its power. Though the vote to confirm Kevin McCarthy, who had led the GOP conference since Paul Ryan’s retirement in 2019, would normally have been pro forma, it turned into a bitter struggle between the HFC and McCarthy’s allies. The caucus demanded McCarthy and the conference leadership take up various fights, such as impeaching Homeland Security secretary Alejandro Mayorkas, and creating a congressional subcommittee on the supposed persecution of January 6 rioters. McCarthy also agreed to give HFC members more committee assignments, to refrain from intervening in primaries without an incumbent, and most consequentially, to bring back the motion to vacate the chair, which allows a single Republican to initiate a new election for Speaker.

McCarthy’s speakership was hamstrung from the beginning by resistance from the HFC. As he attempted to pass budget bills that Joe Biden was willing to sign, and hence avert a government shutdown, the HFC continually denounced him, demanding cuts that would make the bills unacceptable to Biden. As a result, in the spring and the fall McCarthy ended up passing more liberal budgets than he would have preferred, because he had to rely on Democratic votes to pass them. Though the HFC succeeded in weakening McCarthy, their actions made the actual policy outputs of this period more liberal.

In October, the axe fell. HFC member Matt Gaetz filed the motion to vacate, and the House voted 216-210 to remove McCarthy. The race to replace McCarthy was messy. Four candidates tried and failed to win majorities. Mike Johnson, a fourth term congressmember from Louisiana, finally succeeded after a chaotic three weeks. Though not a member of the HFC, Johnson had attended caucus meetings and received money from the group’s PAC. His election was widely viewed as a triumph for the caucus.

Yet while Johnson was quite friendly to the HFC, for the remainder of Biden’s presidency, the group had little success in shaping policy. The Senate still held a slim Democratic majority, so any bill the House hoped to pass needed to be acceptable to Chuck Schumer. When Johnson pushed through appropriations bills that did little to advance HFC priorities, caucus members wailed and gnashed their teeth, but could not change the legislation, and would not try to remove Johnson over the betrayal.

The Republicans’ victory in the 2024 elections returned the party to unified control over the federal government. The margin in the House, however, was quite small. As such, it might be expected that the HFC would yield disproportionate power in this environment. However, the reality has been rather different. As political scientist Frances Lee has shown, “during the first seven months of 2025, the HFC collapsed as a distinct voting bloc in the House Republican conference.” When Trump wanted to pass a continuing resolution that sustained spending levels appropriated under Biden, the HFC unanimously supported him, despite some complaint. Similarly, when Trump’s One Big Beautiful Bill failed to include spending cuts savage enough to offset its tax cuts, the HFC released a document criticizing the bill, and five members voted against a rule to proceed. But when Johnson continued pushing it, the HFC folded, and unanimously voted for the bill they had just described as “morally and fiscally bankrupt.”

As the 2026 midterms approach, the situation of the HFC has deteriorated even further. Almost the entire leadership is leaving the House. In the statehouses, where local franchises of the HFC have been set up, they are losing ground. Wyoming had been their most successful state. The caucus had won majority control of the statehouse in 2024, and announced an ambitious legislative agenda. But after they failed to pass most of it, their reputation sank. In the primaries this week, the caucus lost 13 members in the House and 6 in the Senate.

To sum up, as of this writing, the future of the House Freedom Caucus is uncertain. They do not exist as a distinct voting bloc anymore, and their leaders are departing. If, as seems likely, the Democrats win control of the House in November, their power will be diminished even further, as they return to the status of a minority of a minority.

Where does all of this leave the argument that socialists in office should emulate the caucus’s practice of what Matthew Green has called “legislative hardball”? In my view, there are three things that stand out as particularly relevant from the experience of the HFC.

First, the returns to legislative hardball are highest when the caucus is a minority of a majority, and particularly when there is an opposite-party president. In these situations, the caucus could use its size and discipline to compel the House leadership to enter into fights with the administration that it otherwise would have avoided. These were also the situations in which the caucus succeeded in winning more power for itself in the form of committee apointments, etc.

When the caucus was a minority of a minority, it was effectively powerless and indistinguishable from the rest of its party. When a party majority was combined with a same-party president, the caucus often made noise and complained about policies being made, but was generally unsuccessful in shaping the policy outputs.

Second, the caucus’s actions often resulted in policy moving the opposite direction from what they preferred. Their obstructionism led their party leadership to look for votes on the other side of the aisle in some circumstances. Since some compromises had to be made in order to attract these votes, the resulting legislation was more liberal than it would have been if the caucus had simply supported it. Recalcitrance can have real costs.

Finally, the HFC’s accomplishments have fallen far short of its ambitions. HFC members hoped that their work would build the caucus into a kind of third party in American politics (effectively a right wing version of the party surrogate model that is popular in DSA). This has plainly failed. The caucus has failed to sustain itself as an independent voice in congress, and is deteriorating institutionally as well. The very tactics that gave it its power are in large part responsible for this deterioration. The caucus gained a reputation as obstructionists, and are unpopular with much of the rest of the party.

There are a few implications for socialists in office. On the positive side, the conditions in which socialist legislative hardball are most likely to be successful are those in which the party has a majority in congress and an opposite party president. This is the situation that will likely obtain after the midterms. Given that, I think it makes sense to look for some opportunities for the new socialist bloc in congress to throw its weight around. However, it is worth saying that socialists in congress do not have the kind of organizational coherence the HFC operated with. They do not have an official caucus, or any institutional means of cooperating. Building this should be a priority.

On the whole, however, I think the experience of the HFC should lead socialists to be cautious about legislative hardball as a general stance. The risks are very high. A socialist caucus that rejects legislation as not left wing enough, only to lead the House leadership to look to pass it with Republican votes instead, is not accomplishing very much for the cause of socialism. And obstructionism within the party can actually isolate a caucus, instead of winning a leading role for it.

In other words, I’m afraid the lessons of the House Freedom Caucus for socialists are fairly limited. This is a pity, because the task in which socialist legislators are now engaged, as members of a party they do not lead, is a tremendously difficult one in which there are few models from which to learn. But the experience of the HFC does at least provide some guidance as to the risks of following their example.

Tuesday’s victories were different. In New York’s Seventh Congressional District, Claire Valdez beat Antonio Reynoso handily, despite Reynoso’s position as Brooklyn Borough President, the support of the Working Families Party, and perhaps most importantly, the fervent support of the retiring thirty two year incumbent for that district, Nydia Velazquez. In the Thirteenth CD, Darlializa Avila Chevalier beat Adriano Espaillat, chair of the Congressional Hispanic Caucus and boss of the last remaining machine in New York City politics. Moreover, neither Valdez nor Avila Chevalier are talents on Mamdani’s level. Their wins were powered not by superstar talent, but by the organizing apparatus that NYC-DSA has built over the last decade.

This was not AOC mounting a sneak attack on an unaware Joe Crowley, or Jamaal Bowman beating an Eliot Engel who was too frail to campaign. This was a frontal attack against some of the most entrenched positions in NYC Democratic politics, and it was a total DSA victory (and I’m not even getting into the six out of seven victorious races for the state legislature).

For the Democratic establishment, the announcement of the results Tuesday sounded like one of the seven trumpets heralding the end of days. Staying true to form, the establishment’s response has been to complain about the unfairness of it all. Letitia James mopes that DSA’s candidates “are relatively new to the body politic” (she may wish to reflect on why voters apparently did not view this as disqualifying in the way that she does). Former DNC chair Jamie Harrison asks “If your movement is so strong, why does it keep needing the Democratic Party’s ballot line, infrastructure, volunteers, donors, and voter file to win?”

This line of argument is growing increasingly popular. DSA candidates, it is said, aren't real Democrats, and are using the Democratic Party's ballot line against it. Cuomo dead-ender Melissa DeRosa gave the most vivid articulation of this line of argument when she called DSA “parasites”: “they are going to latch onto the Democratic Party and feed off of it, with the overall goal of taking it over and killing it.”

They aren't wrong.

Parasitism is far from the worst metaphor you could come up with to describe DSA’s relationship to the Democratic Party. But what the establishment doesn't want to admit is that they are ones who have made parasitism a rational strategy. Species, after all, develop evolutionary strategies to fit ecological niches. And the major parties in the US have created a niche for parasites like DSA by making it impossible for third parties to compete.

The most basic way this is true is the use of a single member district first past the post (FPTP) electoral system for all legislative seats, both state and federal. A FPTP system often yields a severe disproportion between votes cast and representation gained. For example, in a district where sixty percent of the population votes Democratic, and forty percent votes Republican, the forty percent of voters who are in the minority don’t have their choice reflected in the outcome at all. The people who cast sixty percent of the ballots are allocated one hundred percent of the seats, since there is only one seat to allocate.

A proportional representation (PR) system leans the other direction. PR requires multi-member districts. So imagine a congressional district elected three representatives to the House, rather than just one. Now, if voting results were the same, the Democrats could conceivably get two representatives, and the Republicans one. Forty percent of the vote would now yield one third of the seats allocated, and sixty percent would yield two thirds. The degree of disproportion between votes and seats would be dramatically decreased.

The disproportion between votes and seats that FPTP systems create makes establishing third parties far more difficult. In a PR system, a party that is not yet established as a major party, and can win only twenty percent of the vote in a given district, can still plausibly win some representation, and thus have something to show for their efforts. In a FPTP system, such a party goes home with nothing on election day, leaving its supporters wondering what the point was. Moreover, in a FPTP system third parties risk playing spoilers, and handing the election to a party they oppose even more than the main party with which they are competing for votes.

In the US, single member districts are mandated for congressional seats by the Uniform Congressional District Act of 1967. The UCDA was enacted primarily to prevent Southern states from reacting to the Voting Rights Act by allocating all of their congressional seats through at-large elections, effectively disenfranchising their new black electorates. In response, congress mandated single member districts as the exclusive form of allocation.1

On the state level, however, parties have more choice. New York state could pass a law tomorrow creating multi-member districts for the state legislature and instituting a PR system to elect them. Indeed, New York City used to use PR to elect members to the city council, but the major parties eliminated it when leftist third parties were able to use it to gain a foothold. And that’s the same reason the Democrats who are now complaining about DSA using their ballot line won’t back a shift to PR today. They want barriers to entry against political competitors.

The problem, of course, is that this system doesn’t actually eliminate political competitors, in the sense of people who don’t like the policy that incumbents are making. Maintaining FPTP just makes it so that starting a third party is a less viable option for accomplishing their goals. But if that road is closed, people will take another. And the obvious alternative is, instead of contesting for power from outside of the dominant party, to instead contest from within.

What’s more, the dominant parties have themselves enhanced outsiders’ ability to contest from within by making primaries mandatory. In 42 states, primaries are compulsory for state level offices. In other words, if your party wants to run someone for governor or state assembly, it must, by law, hold a primary. 86% of the American population lives in a state where this is the case. In the remaining states, primaries are still often used, but there are also options for nomination via convention for some offices.

Combine these compulsory primaries with the FPTP system, and the result is obvious. Challengers to the dominant parties will choose to run in party primaries, rather than trying to build a third party. Winning a primary is, after all, vastly easier. In the general election during a presidential campaign, turnout in New York state runs about 60% of eligible voters. On Tuesday, about 5% of the state’s electorate turned out for the primaries. Winning these elections requires mobilizing a far smaller number of voters. Moreover, once a challenger wins, they basically inherit the party machinery of their opponent. While building a new party requires building institutions for fundraising, communications, etc from scratch, winning a primary allows the challenger to simply take over much of this apparatus from their opponent. When the electoral system creates a niche like this for challengers, it is fanciful to imagine that it won’t be filled. Politics, like nature, abhors a vacuum.

While, in my opinion, both FPTP and compulsory primaries are bad policy, they are different from the next category in that neither is prima facie unfair. I don’t think the state has any business telling political parties how to select their candidates, but if all parties are subject to the same requirement, it’s not really biased against any party. Similarly, FPTP creates strong pressures towards a two party system, but doesn’t really determine which parties will end up on top.

But the US also has a number of laws that simply privilege the major parties and discriminate against third parties. As Seth Ackerman has argued, “some US electoral procedures are unknown outside of dictatorships.” A few states explicitly in state law draw a distinction between the major parties and challengers, and impose restrictions on the latter but not on the former. For example, in Georgia, a political party is recognized by the state if its gubernatorial candidate gets at least 20% of the vote, or if its presidential candidates gets at least 20% nationwide. If a party doesn’t meet this threshold, it needs to collect signatures equal to 1% of the eligible voters in a district to run a candidate. This practice is basically unknown in other rich democracies, and more typical of autocratic rule.

In the rest of the states, the situation isn’t much better. Typically, states require minor parties to gather signatures in order to get on to the ballot. This itself is not unusual internationally. In Canada, for example, candidates for parliament need 100 signatures in a district. What’s distinctive about the US is the scale of signature collecting that is required. In Florida, for example, getting on the ballot (for minor parties, which are defined as parties with less than 3% of eligible voters registered with them statewide; major parties can simply pay a fee to get on the ballot) requires parties to collect signatures equal to 3% of eligible voters in the district. For statewide office, this comes out to more than 400,000 signatures. In general, the Supreme Court has ruled that a signature requirement of up to 5% of eligible voters is not an unfair requirement. In Europe, by contrast, a threshold that exceeds 1% is recognized as a violation of fair election standards.

Moreover, the dominant parties actively change ballot requirements in order to prevent third party competition. In Ohio in 2013, the Libertarian Party was having a lot of success drawing on Tea Party energy and anti-RINO sentiment. In response, Republicans passed a law stripping third parties of their right to participate in state-run primaries, and instituting a requirement (previously recognition had been administrative and basically automatic) of about 50,000 signatures for ballot access. In Arizona around the same time, the GOP passed a law increasing the number of signatures required by a minor party by more than 2,000%, again targeting the Libertarian Party. As a result, the Libertarian Party, which had regularly fielded a dozen or more candidates in Arizona elections, was not even able to get its gubernatorial candidate on the ballot in 2018 or 2022. When the major parties are challenged, they react by making it even harder for third party candidates to compete.

There’s a particular irony to Melissa DeRosa leading the charge against DSA. DeRosa was in the Andrew Cuomo administration, and has devoted her life after his resignation to the quixotic task of restoring his good name (an endeavor in which Andrew Cuomo himself seems considerably less invested). The Cuomo administration actually engaged in exactly the kind of weaponization of ballot restrictions against third parties described above. Here, the target was the Working Families Party (WFP). The WFP is a left wing party that takes advantage of New York’s laws allowing fusion voting, where multiple parties are allowed to endorse the same candidates. Using fusion voting and intervention in Democratic primaries, the WFP managed to build coalition that frequently challenged Cuomo’s reign over state Democratic politics. In response, in 2020 Cuomo pushed through a law targeting the WFP.

Prior to 2021, parties in New York retained ballot access automatically if their last gubernatorial candidate had received 50,000 votes. Under the new law, the threshold moved to whichever was greater - 130,000 votes or 2% of votes cast. and now included presidential votes. A party that declined to nominate a presidential candidate, or failed to meet the threshold, would lose its ballot line. Additionally, the law also raised the number of signatures required for new parties to qualify for ballot access from 15,000 to 45,000.

This is the great irony of the party parasite accusation. DeRosa and other establishment hacks want to argue that DSA is doing something untoward by running in Democratic primaries. But the entire reason DSA has seized on this strategy is that the dominant parties have colluded to make any other strategy impossible.

Parasitism is one of the most common strategies in the animal kingdom. Somewhere on the order of half of all species engage in it for at least part of their lifecycle. It works because there are a lot of ecological niches for parasites to fill. The major parties have created an analogous situation with their cartel-like behavior.

In place of a deteriorating economic situation, Levitz proposes four alternative explanations of college grads’ move leftwards:

1. Demographic change. As the pool of college graduates becomes more female and more non-white, we would expect it to be more liberal, even if “nothing else had changed about America’s society or economy since 1980.”

2. The increasing salience of cultural issues in politics. As issues from gay marriage to DEI policy became more prominent, college graduates, who have always been more liberal on social issues, moved to the left.

3. Coalition pressure. This one works mainly as an extension of 2. As college graduates became more Democratic, their economic views shifted to be more in line with the rest of the coalition.

4. Millenial blues. Millenials are an increasingly large portion of the college-educated pool, and the Great Recession really did injure their economic prospects for the long term. What looks like a secular shift is really a cohort effect.

Here, I want to test Levitz’s first hypothesis, because it’s pretty easy to do so with public microdata. I use three different datasets: the American National Election Survey (ANES), the General Social Survey (GSS), and the Cooperative Election Study (CES). The ANES and the GSS have both been running for more than half a century, while the CES is newer, running only since the turn of the millennium. I’ve excluded 2022 and 2024 from the GSS, since there was a change of sampling mode those years that results in some bizarre weighting assignments and a huge amount of noise.

To operationalize respondent liberalism, I use use the basic liberal-conservative scale that all three surveys deploy. In the ANES and the GSS, it’s a seven point scale, while in the CES, it’s five points. This is sometimes referred to as symbolic ideology, since it inquires about a respondent’s self-identification on an abstract political scale. Another way to approach this would be using what’s called operational ideology, which constructs a left-right scale using a battery of questions about specific issues. Here I’ve gone with symbolic ideology because it is simpler. In all three measures, lower numbers are more liberal, and higher numbers are more conservative.

Looking at the population of college graduates in all three studies, the pattern Levitz describes of a sharp leftwards shift since about 2008 is visible. The ideological shift isn’t enormous, somewhere in the neighborhood of a third of a standard deviation of the scale in all cases. But it’s also hard to miss.

The composition shift in the population of college graduates that Levitz describes is also clear in all three datasets. Since the 1970s, the pool of graduates has grown considerably more non-white, and modestly more female, though that growth has been enough to cross the tipping point into majority female for the entire population (among recent college graduates it’s even more lopsided).

Now, the question is whether the second graph explains the first. The magnitude of change in the second graph is certainly large enough that, if women and non-white people are meaningfully more liberal than men and whites, we can imagine there being relatively little ideological change within each subgroup, but a clear trend in the aggregate as the composition changes.

To examine this possibility, I’ve graphed the mean symbolic ideology across time among white male, white female, non-white male, and non-white female college graduates.

The data with subgroups like this are quite noisy, but the trend is still clear. We don’t see stable trends within subgroups. Instead, each subgroup moves in a more liberal direction after 2008. White men, non-white women, and everybody in between all became more liberal, and the magnitudes of their shift are pretty comparable.

Finally, we can also effectively net out compositional change by holding the composition constant at its average across the entire timespan, and looking at what the aggregate change would have looked like if each subgroup had changed its views in the manner shown in the graph above. If composition effects play a big role, we should see increasing divergence from the observed values as the composition of our simulated study grows more dissimilar to our observed population.

The divergence is pretty minimal, which indicates that composition effects do not play much of a role in the general move left since 2008.

The case for demographic change looks pretty weak. It’s possible that it played a role that this analysis wouldn’t capture, in that white men at college, as they were exposed to more diverse classmates, shifted leftwards as a result. But given that everyone else moved leftwards to roughly the same degree, it seems unlikely that we’re dealing with a bunch of different causes for different subgroups. In light of this, I think we can strike demographic change from our list of suspects.

The last year has brought a great deal of hand-wringing, however, that those younger than millennials (roughly people born after the mid-1990s) are far more conservative. The shift of young voters towards Trump in 2024 is a touchstone for this anxiety. Looking at opinion surveys and exit polls, many worry that future generations in the US will be captured by the right, extinguishing hopes of pushing back the reactionary advance.

I’m somewhat skeptical of this argument, and part of the reason I’m skeptical is that I think it ignores how the millennial political alignment was produced. The experience of the Great Recession and its aftermath was formative for those who entered the job market in those years, and there is strong evidence that it shaped their political beliefs for the long term. And as much as the headline numbers of the US economy may look quite different now than they did in 2008-2015, the actual experience of young workers, particularly those college-educated workers who form the backbone of millennial liberalism, remains more similar to that period than the era preceding the Great Recession.

The US economy just won’t stop producing millennials.

Falling Off the Job Ladder

Over the last five years, some solid research has come out detailing how the worst effects of the Great Recession were really concentrated on the youngest cohort of workers, many of whom were entering the job market for the first time. The most detailed study of this is a recent paper by Kevin Rinz, which estimates cohort level impacts of local increases in unemployment during the recession.

Overall, the national unemployment rate surpassed 10% in the immediate aftermath of the financial crisis. For workers under 25, however, it was nearly double that, reaching 19.5%. For every 1 percentage point unemployment rose in a county, the unemployment rate for millennials rose by 1.2%, over twice the magnitude of any other cohort. In a county whose unemployment rate rose by the average amount, a combination of unemployment and lower wages led the average millennial to lose about 13% of their total earnings from 2007 to 2017, compared to 9% for Generation X and just 7% for Boomers. Moreover, while exposure to higher local unemployment rates stopped showing any effect on Generation X or Boomer wages by 2017, Millennial wages showed little sign of recovery by the beginning of Trump’s first term.

The key mechanism for the wage effect is simple: millennials found themselves crowded out of the jobs at high-paying firms. While from 2005-2010, all cohorts see a sharp decline in their probability of working for a firm whose wages are in the top quartile, after 2010 older generations’ probability begins to improve. Millennials’ does not.

Jesse Rothstein has a paper that zooms in on millennials with college degrees. Using employment scarring models, which estimate the lasting effects of entering a depressed job market, Rothstein finds unemployment rates for college-educated millennials were elevated by the recession, and remained elevated even for labor market entrants in the low unemployment period after 2015. Using estimates of scarring effects from earlier recessions, he finds that the impact of the Great Recession on college educated unemployment is about twice as high as previous models would predict. In other words, compared to earlier recessions, the impact of the Great Recession on millennial job prospects was more severe and longer-lasting.

The Making of the Mamdani Generation

All of this shaped millennial political attitudes. For some time, the canonical paper on the impact of economic downturns on political attitudes was Giuliano and Spilimbergo’s 2014 paper “Growing up in a recession",” which found that early exposure to negative economic shocks increased egalitarian economic preferences. Unfortunately, that paper was retracted in 2023 when the authors announced, somewhat mysteriously, that they no longer had the code or data and were not able to replicate their results.

Fortunately, Jan Bietenbeck and Petra Thiemann recently succeeded in replicating G&S’s key results. Expanding their analysis with further survey waves, and bringing in more international data, they find that experiencing a recession while young increases egalitarian economic values. Recent analysis of UK data finds a similar effect (though interestingly, they find that the introduction of the welfare state after WWII mediates it; generations before the welfare state actually became more right wing in response to recession).

The political effects of the recession made themselves felt quickly. Already by 2017, there were sociologists writing about “a new political generation” and placing millennials at the center of the post-2011 wave of social movements.

Manufacturing Millennials

It would be easy to conclude from this that the forces that shaped millennial politics are not in operation on younger generations. After all, while COVID saw a massive spike in unemployment, “big fiscal” brought that rate down extraordinarily quickly. Moreover, the fact that it was followed by inflation could easily cut the opposite direction, as experiencing inflation tends to make people more conservative.

Yet what this overlooks is something of a revolution in the job market, particularly for recent college graduates. Over the last few months, the plight of these workers has begun receiving some attention, as analysts have noticed their unemployment rates increasing.

The effect is heterogeneous by sex. For female grads, unemployment rates aren’t much higher than average (though as the graph reveals, they remain considerably higher than prior to 2008). For men, however, they are actually nearing the level they reached at the height of the Great Recession. Astoundingly, for young men, the college employment premium has disappeared — college and non-college unemployment levels are the same (though a considerable college wage premium remains).

Rothstein’s argument, it appears, has some staying power. Though headline unemployment is considerably less than during the recession and its aftermath —overall unemployment took until late 2014 to dip below 6%—the job conditions facing recent college grads have remained stuck, and job prospects for the men in this group are already at recessionary levels.

None of this, of course, guarantees that younger generations will follow millennials’ path. In particular, the gender split may well mediate the economic effects, in the same way that black and white workers have quite different political reactions to things like plant closings. Yet I think the evidence points in the other direction. In particular, college educated young men are much less right wing than their non-college counterparts, and are less attitudinally distant from women their age. Indeed, there is plenty of evidence already that the youth swing towards Trump in 2024 was fragile and temporary.

The conditions that shaped millennial political attitudes, particularly for college-educated workers, remain, to a large degree, in effect for younger generations. If they persist, we may well see a decline in age polarization in American politics, as Gen Z and their successors follow the good sense of their elders.

Despite this, I think it’s still fair to say that the notion that the Democratic and Republican parties are much weaker than they were in the past is counterintuitive to most. The parties are obviously massive institutions, and what’s more, in the age of negative polarization (or, in a different language, of hyperpolitics), people are more focused than ever on, at the very least, making sure that the opposing party doesn’t win. As might be expected in light of this, ticket splitting has also sharply declined.

Similarly, third party voting is slowly dying. In the decades after World War II, major third party challenges (receiving more than, say, 5% of the vote) were uncommon but hardly unheard of - the Dixiecrat revolt in 1948, George Wallace in 1968, John Anderson in 1980, and Ross Perot in 1992 and 1996. We’re now approaching three decades without a third party presidential candidate reaching 5%.

In light of this, many find the case for party weakness puzzling. In what sense are the two parties, which are today more dominant than at any other time in American history, weak?

I think the confusion stems from a conflation of the parties with the party system. The parties are institutions, and their strength can be evaluated in much the same way as that of other large bureaucratic institutions - what kind of powers do they have, how autonomous are they, etc. The party system, on the other hand, is a relationship of parties to each other and to potential competitors. The parties are weak, and in large part as a consequence of that weakness, the party system is strong.

Consider American parties in light of the two questions just mentioned. What powers do parties have? How autonomous are they of other institutions? In both of these areas, American parties are quite weak.

The primary function of a political party in a democracy is to run people for office. This is what parties exist to do, and a party that doesn’t do this can hardly be called a party at all (I’m looking at you, Labor Party). How a party chooses who to run for office is thus a core party function. And it’s a party function that, in the United States, is not controlled by parties themselves at all.

In all fifty states, parties are required by law to hold primaries to determine who will appear on the ballot in that state. Parties are simply not allowed to choose their own method for selecting candidates. They do not have that power. As Kinnucan describes, parties cannot choose not to run in a district in order to focus money and attention elsewhere. They are required to hold a primary, and if anyone files the paper work to run in that primary and wins, they are required by law to put that person on the ballot in the general election.

Being required to hold primaries makes the parties gelatinous. They do not have clear divisions separating those in the party and those outside of it. Anyone can run in a party’s primary, and the party is powerless to stop them. Virtually no one in the Republican Party (at least among its elected officials and paid staff) wanted Donald Trump to run in 2016. Very few New York City Democratic party officials wanted Zohran Mamdani to run. In both cases, the candidates ran in large part against what the party as it existed. In both cases, they won, and the party had to change to accommodate itself to them. A system where parties accommodate themselves to candidates, rather than the other way around, is a system of weak parties.

American parties are also weakened by their place in the political ecosystem. As a result of campaign finance law, parties must compete for funds with PACs, super PACs, 501c3s, 501c6s, and various other entities birthed by the tax code. These organizations, which have their own priorities, play an increasingly large role in shaping politics, often to the detriment of parties. In 2024, for example, Republican super PACs took over many of the functions traditionally carried out by the party, like get out the vote mobilization. For the first time, one of the major parties was outspent by super PACs in the campaign. Already diminished by Donald Trump’s personalist rule, which can tolerate no alternative institutional locus of decision-making, the Republican Party’s actual role in the 2024 presidential election was miniscule.

Campaign finance law has also weakened the parties through the requirements it imposes on their fundraising. First, campaign finance law in the US makes the candidate committee, rather than the party, the primary vehicle for fundraising in any given election. This reinforces the candidate-centered, as opposed to party-centered, character of the system. Second, it imposes a financial dependence on the parties that is far greater than that which exists in systems that limit campaign expenditures or fundraising. In Canada, for example, the parties’ total spending for their entire 308 person parliamentary slate is under $30 million. That’s about a third of what Ted Cruz spent on his re-election campaign in 2024. Candidates need huge amounts of money to campaign, which means they need to spend huge amounts of time fundraising. This, in turn, decreases the party’s ability to be autonomous from the sources of those funds.

The result is that American parties are largely unable to act as independent institutions. They’re constrained in their ability to make decisions and perform their core functions. Though the parties excite considerable enthusiasm (though largely negative, from the other side) in the electorate, they exert little control over their own trajectories.

This very weakness, however, is a source of strength for the party system. Precisely because the parties are gelatinous, any challenger to a current party has a better shot of mounting that challenge from within rather than from without.1 In other countries, challengers like Trump or Mamdani would be better off forming a new party rather than attempting to totally reorient an existing one. In this way, the weakness of the parties works to domesticate opposition, who are forced to become part of the coalition they are challenging, if they hope to succeed.

A useful contrast right now is the United Kingdom. There, both the Conservatives and Labour are in free fall, and it’s widely assumed that Nigel Farage’s Reform UK will be the victor in the next election. This is a party that did not even exist at the time of the Brexit vote. Similarly, Zarah Sultana and Jeremy Corbyn have taken steps to found a new party of the left, which already has some half million people signed up, putting it far ahead of any third party effort in the United States of the last quarter century. In the UK, the ability of the Labour and the Tories to control their parties means that challenges have to come from outside.2 The UK has stronger parties than the US, but the party system is looking considerably weaker.

There has been a lot of discussion recently about the unpopularity of the Democratic Party. For some, this unpopularity suggests the time is ripe for a third party challenge. In light of the above, I think this is a mistaken inference. First, the fact that poll respondents are saying they aren’t happy with the Democrats needs to interpreted in light of other responses, like the increase in Democratic identification among voters and the Democrats taking the lead on the generic House ballot. Lots of people are unhappy with the party leadership, but actually there’s a net movement towards the Democrats right now, not away from them.

Then there’s the question of party weakness. Because the Democrats do not control their ballot line, it will always be easier for people who wish to challenge the party’s current leadership to run as Democrats rather than as third party candidates. To do the former, only a bit of paperwork is necessary. To do the latter for any office above city council requires a massive petitioning campaign, and even if that is successful, the prize is a campaign that is all but guaranteed to get less than 10% of the vote.

The twin dynamics of party weakness and party system strength create both challenges and opportunities for those of us on the left. Because the party system is strong, third party politics are not a realistic option, despite the fact that many things would be simpler if socialists could simply run for office as members of the socialist party. Yet the weakness of the parties also creates opportunities. As the Mamdani campaign has forcefully demonstrated, not only is it impossible to exclude socialists from the Democratic Party — it is also quite possible for them to win.

This is a substantial break with the trajectory of American healthcare policy over the last few decades. The Clinton administration launched the Children’s Health Insurance Program in 1997, and today it covers some 7 million kids. George W. Bush enacted Medicare Part D, which expanded Medicare to prescription drug coverage (along with a number of less salutary measures). And of course, Barack Obama signed the Affordable Care Act which, among other things, banned insurance companies from discriminating against patients with pre-existing conditions, gave states a route to expand Medicaid coverage, and launched the ACA marketplaces with subsidies for many families. Largely as a result of the ACA, the number of uninsured people in the United States dropped significantly.

In absolute terms, the number of people without health insurance dropped from about 45 million to about 25 million. Moreover, there were also major changes for people who already had health insurance. If you remember Michael Moore’s Sicko, which came out in 2007, it largely focused on the hardships of people who had health insurance, but were denied coverage via various rules, most notoriously exclusion of coverage for pre-existing conditions, whether or not the insured was aware of the condition. It also required plans to cover some categories of essential care, and banned lifetime benefit limits.

Those of us on the left were quite harsh on the ACA at the time, and with good reason. It was considerably less ambitious than Bill Clinton’s healthcare plan, and the abandonment of the public option eliminated what could have been a crucial beachhead for expanding the public provision of insurance.

At the same time, I think we missed how the ACA ultimately changed healthcare politics. The opportunity it gave states to expand Medicaid was eventually embraced by 41 of them, covering some 20 million additional people. It also meant stories like those in Sicko, of people who thought they had healthcare coverage only to find out in the midst of a catastrophe that they didn’t, would become far less common.

You can see some of the impact of this in public opinion surveys about healthcare. Gallup has asked people for a number of years if they would say they personally worry “a great deal” about the availability and affordability of healthcare. You can see the impact of the ACA in the results.

The shift isn’t huge, somewhere under 10 points, but it’s also unmistakable. And it has taken healthcare from a subject which is of significant uncertainty for nearly a supermajority of the population to one which worries barely more than a majority.

This is a political shift that largely passed the left by, and I think it led many of us to overestimate the resonance of Medicare for All as an engine for Bernie Sanders’ campaigns. To be sure, this demand still found wide support, and polling continues to reveal its popularity (though as Eric Levitz has argued, that popularity can be quite fragile). But the importance of healthcare in American politics has declined steadily as the components of the ACA took effect. Those of us (and I very much count myself in this camp) who thought Medicare for All would be a majority-building platform missed that the demand for this policy had been diminishing. This has been reflected in the broader Democratic Party as well. Since the ACA, Democrats have mainly run on protecting it, rather than significantly expanding it. At the same time, there’s an argument that Republicans have benefited from the low salience of healthcare policy (aside from abortion) over the last decade.

Trump’s budget will change this dramatically. The number of people without insurance is going to rise significantly for the first time since the 2008 financial crisis. For many others, health insurance will get more expensive as subsidies are withdrawn from the ACA marketplaces. Even people whose insurance coverage doesn’t change will deal with the downstream effects of this in hospital closures and related chaos.

In the second chart, you may have noticed that this year, the number of people worrying a great deal about healthcare took a large jump. This is from the survey administered in March of this year. Already by his second month in office, Trump’s administration was driving healthcare worries to their highest level in a decade. As this bill takes effect, that number will in all likelihood continue heading towards pre-ACA numbers.

There’s an opportunity for the left here. This isn’t to say this bill is good for us. It’s a disaster for our priorities in a hundred different ways. But it’s also our responsibility, in its wake, to try and identify where our most favorable terrain is. And there’s just no doubt that the increasing salience of healthcare is an opportunity for the left to renew agitation for one of our most far-reaching and popular priorities.

Subscribe now