The Second Great Realignment

In 1950, America’s leading political scientists published a seminal report called “Toward a More Responsible Two-Party System.” They argued that America’s political parties were — get this — too similar. It seems inconceivable today but was true then. For various historical reasons (mainly the Civil War), the Democratic and Republican Parties both included conservative and liberal members.

In some ways, this ideological incoherence was a feature. For example, it was easier for politicians to traverse party lines when they had cross-cutting loyalties. Party and ideology would frequently come into conflict, opening up a gray area for debate and compromise in which Things Got Done. This had the happy side effect of encouraging ideological moderation.

Still, this arrangement came with too many downsides.

First, in practice, the way Things Got Done usually meant that the interests of Black Americans were sidelined. Conservative Southern Democrats didn’t mind signing bills authored by their liberal Northern counterparts, so long as those bills didn’t challenge Southern racial apartheid.

Second, it was confusing. As the report showed, voters couldn’t tell which party actually represented them. It would be easier if the parties were ideologically sorted. This clarity would lead to better governance, as citizens could easily determine which party was responsible for which policy, and vote accordingly.

Fast-forward through a few major events (mainly Civil Rights) and that’s more or less where we are today. Democrats are liberal, Republicans are conservative, and party brands couldn’t be more distinct.

Call this ideological sorting “The First Great Realignment.”

---

---

This is not a political blog, but politics is an irresistible arena for analyzing the trends sparked by our ongoing technological revolution. I’ve been thinking a lot about the APSA report following Trump’s election victory because I think we’re witnessing a Second Great Realignment, this time precipitated by personalized social media. I also haven’t seen this take elsewhere (possibly because it’s too depressing for Democrats to consider).

Allow me to explain:

Many commentators have remarked (with glee or despondence, depending on their political persuasion) that Black, Latino, and Asian voters swung hard toward Republicans in the last election. What were once reliable Democratic constituencies are no longer so reliable.

People have pinned this on various hobbyhorses (e.g., wokeness) or economic fundamentals (e.g., inflation). But these shifts weren’t a 2024 phenomenon — they’ve been happening slowly but steadily since Trump’s emergence in 2016, with 2024 serving as a kind of “narrative tipping point” due to the scale of the Democratic wipeout:

Something is happening — has been happening — at the tectonic level of our polity.

In particular, non-white Americans are sorting ideologically — just like white Americans long have.

Take Black Americans. Although they’ve long been overwhelmingly Democratic, it’s not because they’re overwhelmingly liberal. If anything, they’re more conservative than white voters. Rather, they have cross-cutting pressures akin to those faced by liberal Republicans and conservative Democrats before the First Great Realignment. There are parts of the Republican platform — potentially salient parts — that Black conservatives might prefer, but overall, they opt for Democrats. That’s partly because they agree with Democrats on other issues, and partly because Black civic institutions encourage norms around Democratic support. Ditto (to a lesser degree) for Asian, Latino, and union-affiliated voters.

Our technological revolution attenuates these cross-cutting pressures in multiple ways.

First, it changes which issues are salient. If border chaos plays well on TikTok (either because the algorithm rewards it or Xi Jinping himself does), then the cross-pressured voters I’m referring to might feel the “conservative” part of their identity activated.

Second, as I’ve written about before, social media has successfully turned many of us into nations of one. We are no longer embedded in vast networks of friends, family, and citizens, meaning civic institutions have a harder time nudging their members to vote in accordance with institutional goals.

You see this most dramatically with unions, who could once ensure that even their most culturally conservative members voted Democrat. That discipline has broken down, so much so that they wouldn’t even support Biden, the most aggressively pro-union president in living memory.

It would be judgmental of me to say that union members are voting against their interests. Many of them are simply conservative! But it used to be that unions could influence their conservative members to vote based on economic factors that favored Democrats. Now, however, union leaders feel sheepish even endorsing Democrats. Technology has uncovered the “true preferences” of their constituents, and leaders would rather not lead by explaining why these preferences might be suboptimal. (By the way, the same thing happened with the Republican Party as a whole, which Ben Thompson discusses here. TL;DR: Republican leaders tried pushing back against the Trump phenomenon, but there was no arguing with their base, who developed a deep and visceral attachment to him. Eight years on, the Republican Party is Trump’s party.)

Moreover, as we’ve replaced human interaction with a technological facsimile, the connective tissue binding communities together has eroded. Even if a union member was culturally conservative, they knew that their vote meant something beyond themselves. They had a communal obligation to vote for the betterment of their peers. (I make no normative claim as to whether the outcomes of such obligations were good or bad.)

Anyone who’s driven on a freeway recently knows that communal obligations are dead.

Finally, and most importantly, social media has given voters better information about which of the two parties better represents them. (This position might be controversial.) It’s easier than ever to “match” one’s political preferences to Democrats or Republicans. Personalized algorithms ensure it.

Every group — whether it’s young people or elderly Latinos — includes a mix of liberals, conservatives, and moderates, but now they’re completely capable of self-sorting into the “right” ideological bucket. This isn’t because the parties have clarified themselves, as in the First Great Realignment, but because voters, empowered by social media and niche information sources (whether Joe Rogan or Call Her Daddy), develop a particularized attachment to a way of thinking about the world, a “vibe.” To the extent political preferences are downstream from biologically innate moral intuitions, this effect is all the stronger. In the same way social media can suss out your deepest desires, it can determine whether you are genetically predisposed to valuing orderliness (conservative-coded) or fairness (liberal-coded).

In short, we all vote with our lizard brain now.

I feel for the Democrats. Their strategy was predicated on a generational head fake: They thought they could hold on to non-white voters in a diversifying nation while young people became increasingly progressive. But it looks like Millennials were a one-off when it came to progressivism, which I’d chalk up to the fact that we came of age in the shadow of George W. Bush’s catastrophic presidency. The disasters of the Iraq War, Hurricane Katrina, and the Great Recession into which we graduated more or less guaranteed we’d skew liberal. Gen Z’s relatively even Republican/Democratic split represents a return to normal. Even the gender divide that seemingly caught everyone off guard isn’t surprising: Men are more conservative than women, and again, the internet empowers them with the information they need to vote accordingly. (Of course, algorithmic incentives likely exacerbate these differences by cordoning off men from women in separate informational ecosystems. As I like to say, when it comes to analyzing any sufficiently complex social phenomenon, the x-axis is the y-axis is the x-axis.)

---

I want to acknowledge that I could be wrong: Even though the realignment I’m gesturing at has been an ongoing process, it could also be a fluke, a one-off, brought on by a pandemic, inflation, and a border crisis. Non-white voters were disproportionately affected by all of these issues, which could have prompted them to punish the Democrats without any great love for Republicans. My thoughts could look silly by 2028, especially if Trump tries to reinstitute the Know-Nothing Party while allowing Elon to boss him around.

I also don’t want to underrate the effects of particular personalities or strategic choices taken by both parties. Trump is a unique figure, perfectly suited to our technological moment. (What this says about our technological moment I leave as an exercise to the reader.) His brand is over-the-top; his soundbites are brash. He’s so unsubtle that he essentially forces voters to re-arrange themselves into pro- and anti-Trump camps, a process facilitated by hyper-efficient algorithmic sorting. Had the Republicans succeeded in stopping Trump’s rise, the Second Great Realignment might still have happened, but more slowly. It’s hard to imagine the algorithm getting a lot of mileage out of Jeb Bush.

Similarly, if the Democrats hadn’t tried to sneak an aging Biden past a suspicious populace, we might be celebrating the election of President Whitmer. History is always contingent.

Still, I think it’s worth grappling with this thesis because even if it’s at least partially correct, the implications are profound:

1. Democrats will need to moderate: I know this will upset some folks, especially ones who don’t view the Democratic Party as potentially progressive outside cultural issues, but the conclusion is unavoidable. If voters are increasingly sorting based on ideology instead of identity, Democrats will need to fight hard to attract more moderate defectors, white and non-white alike.

2. It’s never been a better time to be a billionaire: Another way to summarize my post is that “capital has won.” Working class Americans increasingly identify with a party whose stated policy goals — tariffs, DOGE-style spending cuts, etc. — would actively harm them. Meanwhile, Democrats have become the party of economic winners — people who are so insulated from material concerns that they might (might!) favor intellectual purity over the messy compromises it would take to win working class voters back to their side.

3. The Biden Administration represents the high-water mark for progressivism, at least in the near-term: I hate this conclusion. The economist Tyler Cowen has been promulgating it on Marginal Revolution for a while, and I still hate it. We have several problems — particularly housing and the environment — that will require a revitalized government to help solve. But in sketching out my argument, I’ve come to believe Cowen is right. The Inflation Reduction Act, if it survives, is the best progressives can hope for in the near-term.

There is, possibly, a fourth conclusion, which I’d characterize as “Bernie Part 2.” The Democrats could, like Trump, try to activate highly disengaged and marginal voters with a more aggressive populist message. It didn’t work in 2016 or 2020 because the Democratic Party and its affiliated media could still exercise enough institutional control to elevate Clinton and Biden. But, as we’ve shown, the rules of the game have changed in a major way. If Democrats try to moderate and fail, it opens up a pathway for AOC.

Regardless of which path Democrats take (which will no doubt depend on the strategic choices Republicans take in a second Trump Administration), it will be a turbulent four years.

Such is always the case with realignments.