Midterm reckoning: democracy's stress test

Americans go to the polls today in what every pundit is calling the most consequential midterm election in a generation — a phrase deployed so reliably every two years that it has lost nearly all meaning. Yet the circumstances of November 2026 are genuinely unusual: a sitting president whose approval rating stands near a record low on the eve of the vote, a Supreme Court that this week rejected an emergency White House request to set aside 20,000 California ballots before counting began, and a campaign season that has shattered all spending records to become the most expensive in American history. The machinery of democratic accountability will today receive one of its more rigorous stress tests in living memory.

The received wisdom

The mainstream framing is straightforward enough: the midterms are a referendum on Trumpism, and the numbers look bad for the Republicans. Polls show the president well underwater nationally, inflation remains the dominant voter concern, and the spectacle of the administration attempting to pressure election officials — before being ordered to stop by a federal judge — has reinforced the narrative of a White House willing to bend democratic norms to stay in power. The California ballot episode in particular drew bipartisan condemnation: the Supreme Court, including justices appointed by Trump himself, unanimously declined to intervene. For mainstream commentators, this is the story of institutions holding — just.

On the Democratic side, there is cautious optimism. Progressives aligned with New York Mayor Zohran Mamdani are fielding candidates for Congress, and enthusiasm among younger voters remains elevated. The party sees an opportunity to win back the House and, in some projections, hold the Senate. The economic data, however, offers a more complicated backdrop: petrol prices have dipped slightly ahead of the vote, a suspiciously convenient softening that the White House will not be shy about claiming credit for.

A different read

Step back from the daily news cycle and this midterm looks less like a clean referendum on any one presidency than like a symptom of a deeper structural crisis in American democratic life.

Begin with the money. The most expensive midterm campaign in history did not get that way because Americans care more about civic life. It got that way because both parties have become extraordinarily efficient at converting outrage into donations, and because the post-Citizens United landscape has made electoral spending a near-limitless arena for billionaire intervention. The shattering of spending records is not a sign of a healthy democracy — it is a sign of one that has outsourced its energy to institutional donors rather than citizens. The right spent decades warning about campaign finance distortions and then largely abandoned the critique once it found itself on the winning end of the donor ledger. That was a mistake, and the present situation is partly the consequence.

Then there is the AI problem. A new NPR investigation finds that AI chatbots are increasingly being used as political advisers by voters seeking information about candidates, polling places, and ballot measures — and that their answers are frequently wrong, misleading, or subtly skewed. This is a genuinely new and underappreciated threat to informed voting, distinct from the social media disinformation debates of the past decade. A voter who consults a chatbot and receives a confident but incorrect answer about registration deadlines is not a victim of malice but of epistemic infrastructure that was simply not designed for this purpose. The research on political polarisation suggests Americans are actually less ideologically sorted in their daily lives than the media coverage implies — but they are making decisions in an information environment increasingly hostile to the kind of slow, contextual reasoning that good civic participation requires.

The ballot challenge episode deserves particular attention, because it is being misread by both sides. Conservatives who cheered the attempt are wrong: there is no coherent principle of election integrity that supports asking a court to sequester valid ballots on the eve of counting. But liberals who see the Supreme Court’s rejection as proof of institutional resilience should be more cautious. The court did not rule on the merits of any broad election claim; it declined an emergency application, as courts routinely do when the relief sought is drastic and the factual record is thin. The legal question of DOJ’s role in election pressure — where a judge had to order the administration to stop threatening officials — is more constitutionally serious, and it received less coverage than the ballot story.

What is most striking about the Mamdani-aligned progressive candidates running for Congress today is how little their candidacies have been covered in terms of concrete policy. The left-progressive movement that produced Mamdani’s New York victory has genuine grassroots energy, but its national agenda — on housing, on trade, on defence spending — remains strikingly vague for a movement seeking to reshape federal legislation. Enthusiasm is not a policy platform. The Democratic Party’s fundamental tension, between its institutional donor base and its activist base, will not be resolved by a midterm result no matter which way it goes.

History offers a useful reminder: the 2006 midterms, which gave Democrats a sweeping House majority, ushered in a period of intense legislative ambition that produced the Affordable Care Act — and then the Tea Party backlash of 2010. Pendulum swings in American politics tend to overshoot in both directions, and the structural conditions that produce them — economic anxiety, cultural resentment, institutional mistrust — are not cured by electoral outcomes. They require the harder, slower work of governance.

What to watch

  • House control: Whether Democrats flip the chamber will determine the trajectory of the next two years more than any Senate race.
  • California: The ballot challenge, the voter fraud charges against six people filed by Trump’s DOJ, and the state’s sheer electoral scale make it the night’s most consequential battleground.
  • The Mamdani effect: Whether his endorsed candidates outperform or underperform will indicate how transferable his coalition is outside New York’s specific conditions.
  • Post-election legal challenges: Given the pre-vote litigation activity, the question of whether the losing side accepts results quickly or contests them through the courts will matter enormously for institutional confidence.

— J