"Default settings" implies the pre-2016 configuration was an equilibrium rather than a way station. The evidence points the other way. The educational realignment that defines current coalitions was underway roughly two decades before 2016 — researchers studying the flip in the education-vote relationship describe it as a trend spanning elections rather than one triggered by a critical election, and CES data suggest the realignment was already in motion before Trump ran. It is also not an American phenomenon: Grossmann and Hopkins argue the same trends are occurring globally, with the U.S. distinguished mainly by a two-party system that forces populist energy inside an existing party rather than into a new one. Canada's 2025 election showed the same pattern — gains among university graduates, losses among voters without degrees.

Worth flagging a live empirical dispute: some political scientists argue the diploma divide largely disappears once you control for race and geography, and that what looks like a class realignment is a regional and racial one. That debate isn't settled.

So the honest framing isn't "will America revert," it's "which changes are cyclical and which are ratchets."

Trends that run independent of Trump

Coalitional divergence by gender among the young. This is the clearest genuinely new development, and it isn't a reaction to any policy. Pew's National Public Opinion Reference Survey found the Republican identification gap between young men and young women grew from 6 points in 2020 to 22 points, with roughly half of Gen Z men identifying Republican against 27% of Gen Z women. Catalist measured a 17-point gap in 2024 — the largest across four presidential cycles. The mechanism runs through media consumption, family formation, and labor-market position, not through any particular administration. Burge notes that marriage and childbirth — historically a moderating force — are themselves sorting by partisanship, with 71% of young conservative women reporting parenthood against 40% of liberal women aged 25–35. If that holds, it compounds demographically rather than reverting.

Conditional institutional trust. Gallup's confidence average has sat below 30% for five straight years, and the agency's own read is that Americans no longer share broad faith in civic institutions — confidence now tracks who holds power. Past declines have reversed, but each recovery has landed lower than the last. That's a ratchet description, not a cycle description. It also predates 2016 by decades.

Executive power accretion. This is the strongest structural case, and it's explicitly bipartisan. AEI's framing is that each presidency's novel step becomes the next one's baseline, that both parties object when out of power and rely on the precedent when in it, and that the next Democratic administration should be expected to extend rather than surrender these innovations. Democrats are saying so on the record — AOC told NOTUS that correcting current abuses requires acting in similar capacities. And the legal architecture is now locked in: in Trump v. Slaughter (June 29, 2026), the Court held 6–3 that for-cause removal protections for FTC commissioners violate the separation of powers, overruling Humphrey's Executor and casting doubt on protections across commission-style agencies. A future president of either party inherits that.

The tariff case cuts the same way even though it went against the administration. The Court held on February 20, 2026 that IEEPA doesn't authorize presidential tariffs — but new tariffs issued days later under Section 122 of the Trade Act, and Section 232 and 301 tariffs on steel, aluminum, copper, autos, lumber, trucks and semiconductors were untouched. The instrument changed; the policy didn't. Inference on my part: that substitution pattern is what a durable regime change looks like as opposed to a personality-driven one.

Procedural escalation as a norm. Six states enacted new congressional maps between the 2024 and 2026 elections — the highest level of mid-decade redistricting since the 1800s. Before 2025, only two states had done voluntary mid-decade redistricting since 1970. The states that did it dismantled independent commissions to do so. Undoing that requires each side to unilaterally disarm.

The information environment. About one in five Americans regularly get news from social-media influencers, rising to 38% of adults under 30, and roughly three-quarters of those influencers have never worked in a newsroom. Reuters Institute finds news interest has fallen about 15 points across five years. This changes who can be a political entrepreneur and lowers the barrier for insurgent candidacies in both parties permanently.

Party organizational hollowing. Michael Kazin's assessment of the current primary environment: formal party organizations are an empty shell, so outside groups with money and supporters now constitute the party. That's a decades-long trend accelerated by small-dollar fundraising and super PACs, and it makes both parties less able to enforce discipline — which is precisely what made a hostile takeover possible in the first place.

The fiscal constraint. This one binds regardless of who wins anything. CBO projects nominal interest costs more than doubling from $970 billion in 2025 to $2.1 trillion by 2036, with debt reaching 120% of GDP. On the 30-year path, net interest becomes the single largest federal expenditure, surpassing Social Security by 2048. Any post-Trump administration of either party inherits a policy space narrowed by debt service. Inference: this makes both parties more likely to reach for regulatory and executive tools, since those don't require appropriations.

A new cleavage forming. AI infrastructure is emerging as an issue that doesn't map to existing coalitions. An Annenberg poll found 61% opposed to new data centers locally — 69% of Democrats and 54% of Republicans. Local opposition blocked or delayed 75 projects worth $130 billion in the first quarter of 2026 alone. Eight states rolled back data center tax subsidies in 2026. Cost-of-living, energy prices, and concentrated tech power are being fused into a single frame by figures on both flanks.

The strongest counterargument

Thermostatic public opinion is real and currently visible. Gallup's June 2026 reading found 79% saying immigration is good for the country and only 30% wanting it reduced, down from 55% in 2024 — with the largest movement among Republicans, whose support for cutting immigration fell 40 points to 48%. That is the American electorate behaving exactly as the "default settings" thesis predicts: pushing back against whatever direction policy is moving. The generic ballot sits around D+6 as of late August.

And the coalition itself may not survive its assembler. One argument making the rounds is that MAGA's members had little in common beyond a favorable view of the project — hawks and isolationists, free marketeers and protectionists, deficit hawks and pro-entitlement populists — and that without a unifying figure only one side's vision can win. Ted Cruz is already arguing publicly for a return to economic conservatism and away from trade populism.

Where I'd land

Policy direction is cyclical; institutional form is not. Expect substantial reversion on immigration levels, regulatory posture, and personnel — the thermostat is already running. Expect no reversion on: presidential removal authority, emergency-power practice, mid-decade redistricting, the education-and-gender structure of the coalitions, conditional institutional trust, or the hollowed party apparatus. Those aren't Trump artifacts; several predate him and all now have bipartisan constituencies.

One more piece of context: 2024 was the first year since records began in 1905 in which every incumbent party in a developed country lost vote share. Whatever is happening is not primarily an American story about one man.