Forget the 2026 ‘Wave.’ Forecast the House District by District.

Ask the wrong question and you'll get the wrong map. The 2026 midterms are not one national election. They are 435 separate constitutional contests held on the same day under one Constitution. The pundit class wants to forecast a wave. The Constitution wants you to forecast districts.

So here is the angle most national forecasters miss. The proper unit of analysis for 2026 is not the national mood. It is the district. And forecasting districts is a Hayek problem, not a polling problem.

The Constitution Has No Generic Ballot

Article I, Section 2 says House members are chosen every second year by the people of the several states. Not by the nation. Not by a national generic ballot. James Madison wrote in Federalist 52 that the House should have an immediate dependence on, and an intimate sympathy with, the people. That is a local, district-level standard.

A national poll is a political abstraction. A House seat is a legal and geographic fact. The originalist approach to midterm forecasting begins with that distinction.

Think about what the generic ballot actually does. It asks a nationally representative sample whether they prefer a generic Republican or a generic Democrat. But no voter anywhere casts a ballot for a generic anything. They cast a ballot for a named human being standing in a specific county, city, or township, before a specific set of neighbors. A swing seat in suburban Phoenix is not the same contest as a swing seat in northeast Pennsylvania. The Constitution never asked them to be.

The Knowledge Problem Applies to Elections Too

Friedrich Hayek's central insight in "The Use of Knowledge in Society" was that the information required to run an economy is dispersed among millions of individuals and can never be collected in a single central office. The same is true of a 435-seat House election.

The information that decides a district race-the candidate's local reputation, the factory that just closed, the school board fight, the county GOP chairman's field operation, the church community's turnout, the ballot measure driving voters-exists only in fragments. It cannot be compressed into one national number without losing the very thing that decides the election.

The generic ballot is central planning applied to politics. It pretends a remote analyst can know what 435 local electorates will do better than the voters themselves. That is the same conceit that fails in economics, and it fails in election forecasting. You cannot plan a national House majority from a Washington newsroom any more than you can plan steel production from a federal ministry.

This is not a minor academic point. It has real consequences. A campaign built on a nationalized cable-news narrative is a campaign that ignores the price signals of local politics. And local political information is the only information that actually moves a House seat.

The Seen and the Unseen in the 2026 Map

Henry Hazlitt taught us to look not merely at the immediate effect for one group, but at the longer effects for all groups-the seen and the unseen. National fundamentals are the seen. The district-level machinery that decides close races is the unseen.

The visible national story in 2026 will be about the president's party and the midterm penalty. That is real. The president's party almost always loses House seats in a midterm. But that is headwind, not forecast. The actual majority will be decided by the unseen factors that no national model captures well.

The decisive 2026 forecast will come from tracing the consequences of several district-level forces, not from a single national number.

Crossover seats are the first force. These are districts that split between the presidential and House results in 2024. They are the true battlegrounds, where local candidate quality still outweighs national trend. A district that voted for one party at the top of the ticket and the other party for Congress is telling you that a national wave misses the point.

Open seats and retirements are another. Candidate recruitment matters more than any national poll. An open seat in a marginal district can flip on a weak filing deadline. One bad recruit in a swing seat is worth more than a two-point shift in the generic ballot.

Redistricting litigation is still unresolved in several states. One court ruling can shift three or four seats before a single vote is cast. Until those lines are final, any national seat projection is guesswork with lawyer fees.

State ballot measures also change the equation. School choice, parental rights, crime, energy, election integrity. These do not show up in the generic ballot, but they can drive district-level turnout in ways no national model captures. A ballot measure can change the composition of the electorate as much as any candidate can.

Then there are local price signals. Housing costs, gasoline prices, grocery bills. Voters experience inflation locally, not as a national CPI release.

The smart 2026 analyst will not ask what the national generic ballot says. He will ask which party can recruit a credible candidate in the 30 districts that actually decide control.

Prices Are Local, and So Is the Cost-of-Living Vote

National inflation is an average. But a voter in suburban Phoenix experiences housing costs differently from a voter in rural Pennsylvania or a Gulf Coast energy district. The price system sends local signals, and district-level forecasting should read them.

This is Thomas Sowell's incentives at work. People vote their lived circumstances. A district where housing costs rose 20 percent is a fundamentally different race from one where gasoline prices dominate. A national inflation hurts incumbents narrative is too blunt. The district map will punish some incumbents badly and leave others untouched.

Take a hypothetical but typical suburban district outside Phoenix. Housing costs there are not a national statistic; they are a monthly rent increase, a property tax bill, a decision to move farther out. The voters in that district do not need a Federal Reserve press release to know prices rose. They see it in the strip mall and the grocery aisle. That lived reality shapes turnout and persuasion in a way no national poll can model.

Now take a district in rural Pennsylvania where the dominant economic concern is energy costs and the local permit process. The issues are different, the candidates are different, and the winning message is different. Both districts might show up as inflation concern in a national poll, but the campaigns that win them will be local campaigns aimed at local conditions.

This is why the central-planning model of midterm forecasting fails. It aggregates things that cannot be aggregated without destroying the local knowledge that matters.

Where That Leaves the 2026 Forecast

Republicans enter 2026 with a razor-thin House majority. The president's party historically loses ground in midterms. That is a headwind, not a forecast. The actual majority will be decided by a few dozen districts, many by a few thousand votes.

Anyone who hands you a precise seat count now is doing the central-planning thing. I will not do it. The data that matters-candidate filings, primary results, district-level turnout, court-drawn lines-does not exist yet.

But the structural forecast is clear.

If Republicans run a district-level campaign-recruiting candidates who fit their districts, running on local prices, local crime, and local schools-they can hold the House despite the midterm headwind. If they run a nationalized campaign built around cable-news narratives, Democrats can flip enough crossover and open seats to take the majority.

The real battleground is not the national mood. It is the candidate filing deadline in 30 districts.

Think of it this way. A national wave is what forecasters see from 30,000 feet. A district-level campaign is what voters experience on the ground. The Constitution was written for the ground.

The Republic Is Not a Plebiscite

The framers gave us districts so that representatives would answer to their neighbors, not to a national narrative. The House is the people's chamber, and it is supposed to be local.

That is not an accident. It is a structural safeguard. A pure national plebiscite would flatten local differences and concentrate power in the loudest national voices. The district system preserves the dispersed knowledge of a free people. It forces candidates to listen to the actual voters in front of them, not the abstract electorate imagined by a pollster.

A district-level forecast is not just better political science. It is the originalist, constitutional way to think about 2026.

The national pundit wants a wave. The Constitution wants a map.

Forecast the map.