The election did not move much. The world around it did.

One week before the November 3 midterms, the most striking thing about the national polling is how little it has changed. The RealClearPolitics generic-ballot average stood at Democrats 50.5% and Republicans 42.4% through October 20, a Democratic advantage of 8.1 points. Silver Bulletin's October 23 average was D+9.1, almost exactly its cycle high. The individual polls differ, sometimes substantially, but the broad national environment has been sitting in roughly the same place for several weeks.

That would make for a rather dull update if elections were only about polling averages. They are not. During the past week, several forces that had been running alongside the campaign have moved directly into it. The war with Iran is now showing up at the gas pump and in household budgets. The Justice Department has intervened in election administration in California and filed a pre-election lawsuit over the Alaska Senate contest. Early voting has accelerated. And voters are telling pollsters that the next Congress should spend less time replaying the partisan argument and more time lowering costs and ending the war.

In our October 19 update, the central finding was that the map had moved more than the national vote. That still looks right. What has changed in the final week is the nature of the trust problem. It is no longer confined to whether voters trust one party more than the other to govern. It increasingly includes whether they trust institutions to administer elections, whether international decisions are being made with domestic consequences in mind, and whether the eventual winners understand what voters actually want them to do once the campaign is over.

This is a more complicated election than it was six days ago, even if the topline polls look familiar.

The economy has absorbed foreign policy

Foreign policy rarely decides a midterm election by itself. This year it may not have to. The international story has been translated into a domestic one.

Reuters reported on October 24 that the war with Iran and the administration's tariff policies are increasingly being understood by voters through their effect on prices, especially energy costs. In a Reuters/Ipsos poll, 78% of Americans said they blamed President Trump's policies for prices rising faster than their earnings. The same survey series has placed his overall approval at roughly one-third of the public, while separate AP-NORC polling earlier this month found only 17% of adults approving of his handling of the cost of living and 26% approving of his handling of the economy.

The most revealing new number, however, may be about priorities rather than approval. An AP-NORC survey released October 22 found roughly 80% of registered voters saying the next Congress should make reducing the cost of living a top priority and about 70% saying it should prioritize ending the Iran conflict. Around half prioritized reducing the federal deficit and finding bipartisan common ground. About 40% wanted investigations of the Trump administration to be a top priority, while only about 20% wanted Congress primarily to advance Trump's agenda.

That combination is revealing because it collapses the old distinction between 'domestic' and 'foreign' policy. For many voters, Iran is now an economic issue. Tariffs are an economic issue. Energy policy is an economic issue. The question is no longer whether voters are following international affairs closely. It is whether they believe decisions made abroad are raising their costs at home.

From an Optimal Trust perspective, this is an Aligned Interests problem as much as a Competency problem. Voters can believe that a government is acting decisively overseas and still conclude that the consequences are misaligned with their own economic security. A policy may be coherent on its own terms and still lose trust if the public experiences it as something being done to them rather than for them.

The Justice Department is now part of the trust story

The most unusual development of the week may be the Justice Department's entry into election administration itself. It deserves careful treatment because the facts are serious enough without embellishment.

In California, the Justice Department sent letters to all 58 counties demanding help verifying the citizenship of people on federal lists of suspected noncitizen voters. The department asked officials to segregate mail ballots associated with people whose citizenship could not be confirmed and imposed a 48-hour response period. California election officials have challenged the accuracy of the lists and the federal government's authority to demand the information. The Justice Department says the effort is intended to protect election integrity and is not a request for mass voter-registration purges.

The details matter. According to Associated Press reporting, Los Angeles County reviewed an initial sample of 200 names and found that more than 40% had already canceled their registrations, more than 70% had no voting history, and nearly 40% were U.S.-born. Those findings do not by themselves resolve the broader dispute, but they show why the reliability of the federal lists has become part of the controversy.

Alaska raises a different and potentially more consequential question. On October 23, the Justice Department filed suit seeking to prevent certification of the Senate election if Republican incumbent Dan Sullivan loses by fewer votes than are received by another candidate named Daniel J. Sullivan Jr. The department argues that the second Sullivan's presence could confuse voters and distort the result. President Trump subsequently said that he personally directed the administration to file the suit. The incumbent senator's campaign has said Alaskans can distinguish between the two names, while allowing the case to proceed. Mary Peltola and Democratic groups are seeking dismissal.

These are legal disputes, and the courts will decide the legal questions. The political significance is different. The Justice Department is no longer merely an institution whose conduct might be investigated after the election; it has become an actor in disputes about how votes are handled before the election and how a closely contested Senate result might be certified afterward.

That pushes the election into the Systemic level of the trust architecture. The central question becomes not simply whether voters believe their own ballots will be counted, but whether they believe the rules will remain stable and institutionally neutral when control of Congress is at stake. In our October 19 article, we noted evidence that confidence in the mechanics of ballot counting may actually be stronger than some earlier surveys suggested. That finding still stands. But the institutional environment around the count has become more contested. Systemic Rational trust - confidence in procedures - and Systemic Emotional trust - confidence that the system will be accepted as legitimate - are being pulled in different directions.

The Senate is no longer a short list

The Senate map continues to be the part of the election where national dissatisfaction meets candidate-level politics most directly. Cook Political Report's October 22 ratings list seven Senate races as Toss Ups: Alaska, Iowa, Kansas, Maine, Michigan, Ohio and Texas. North Carolina and Georgia are in Likely Democratic, Minnesota and New Hampshire in Lean Democratic, South Carolina in Lean Republican, and Mississippi and Nebraska in Likely Republican. Those are Cook's ratings, not an Optimal Trust forecast, but the distribution shows how wide the competitive field has become.

The latest polling reinforces the same point while also warning against false precision. Recent public polls have shown Sherrod Brown ahead of Jon Husted in Ohio by margins ranging from three to five points; Troy Jackson and Susan Collins tied in one October 23 Maine poll; Texas polls ranging from a tie to small leads for James Talarico; and Alaska polls with Mary Peltola and Dan Sullivan separated by only a few points. Iowa has produced contradictory late polling, with some surveys showing Josh Turek slightly ahead and others Ashley Hinson slightly ahead. These are precisely the kinds of margins that should be read as competitiveness rather than destiny.

North Carolina looks different. Recent polling has repeatedly shown Roy Cooper with a substantial lead over Michael Whatley, including 52%-39% in an Elon survey released October 20. Georgia also appears less volatile than the Toss Up states in recent public polling. But the Senate majority is determined by the collection of state outcomes, not by a national popular vote, which is why a broad Democratic national advantage does not translate mechanically into Senate control.

The deeper OT question is whether the national environment is overwhelming candidate-specific trust or merely creating the conditions in which it matters more. Maine is a useful example: Susan Collins has spent decades building an individual identity distinct from the national Republican brand. Texas poses the opposite challenge, where the national environment and a highly polarized candidate contest are tightly intertwined. Alaska now adds a systemic dimension because the dispute over the ballot and certification is occurring inside an already close race.

The final week therefore should not be read as one national wave moving uniformly across fifty states. It is better understood as a strong national pressure being filtered through very different local trust structures.

The House story is still the map, not the headline number

The House remains simpler in one respect and more complicated in another. The generic ballot continues to show a substantial Democratic national advantage, but Cook's most recent House ratings summary, updated October 16, still places 214 seats in Democratic-favored categories, 200 in Republican-favored categories and 21 as Toss Ups. There has not been a later Cook House-wide ratings update through this morning, so the responsible approach is not to manufacture a new seat count from a handful of polls.

What we can say is that the national environment has not snapped back toward Republicans during the past week. RealClearPolitics' average is D+8.1; Silver Bulletin's October 23 average is D+9.1; the October 19 Reuters/Ipsos poll was D+12; and the New York Times/Siena national survey reported a 10-point Democratic lead when respondents were shown the actual congressional candidates in their districts. Those are different methodologies and should not be averaged casually, but they point in the same broad direction.

Siena's survey also helps explain the persistence of that environment. President Trump's job approval was 36% approve to 59% disapprove, and he was underwater on every major issue tested, particularly cost of living and gas prices. At the same time, the poll found broad dissatisfaction with both political parties. Voters gave Democrats comparative advantages on the economy, cost of living, health care, foreign wars and several other issues, while majorities still believed both parties were focused on the wrong things.

That is almost a textbook example of the Trust Transfer Gap we identified on October 4. A comparative advantage is not the same thing as institutional affection. A voter can conclude that one party is currently better positioned to address an issue without concluding that the party itself deserves deep or durable confidence. The ballot records the choice. It does not record the reservation attached to it.

Twenty-six million votes have already been cast - but do not turn them into a forecast

The campaign is also no longer entirely prospective. By October 25, roughly 26 million ballots had already been cast across 46 reporting states, according to an early-vote tracker that compiles state election-office data. In states that report party registration, registered Democrats have accounted for a substantially larger share of ballots than registered Republicans, and early turnout in several competitive states has been running ahead of comparable points in 2022.

That is important evidence of engagement. It is not a reliable prediction of the final result. Early voters differ from Election Day voters, states begin early voting on different schedules, party registration is not the same as vote choice, and campaigns deliberately shift supporters between voting modes. The lesson from the early vote is therefore narrower: the enthusiasm we identified in the October 19 article has now moved from survey responses into actual participation by millions of people.

That matters because one of our original questions was whether dissatisfaction would produce withdrawal or action. At least for a substantial part of the electorate, it is producing action. The remaining question is how balanced that mobilization becomes as Republican campaigns concentrate on late early voting and Election Day turnout.

What changed since October 19?

The best way to understand the past six days is not as another directional swing in the polling. It is as a widening of the election's causal field.

The first change is that international events have become domestic economic events. The war with Iran and tariff policy are being judged through gasoline prices, household costs and perceptions of economic security. The second is that election administration itself has become a live campaign issue through federal legal action in California and Alaska. The third is that turnout is becoming observable rather than hypothetical: tens of millions of votes are already in. And the fourth is that voters are being unusually explicit about what they want after the election - affordability and an end to the Iran conflict rank far above either partisan investigations or simply advancing the president's agenda.

What has not changed is equally important. The national generic ballot remains broadly where it was. Voters remain unhappy with both parties. The Senate remains highly dependent on candidate and state context. And the central Optimal Trust proposition still looks useful: withdrawal of trust from one governing coalition does not automatically create equivalent trust in the alternative.

Indeed, the final week may be strengthening that proposition. If voters are primarily asking for lower costs, an end to war and a government that functions without constant escalation, then even a decisive electoral result would not necessarily amount to a mandate for a broad ideological program. It could instead be a demand for correction. Those are very different forms of political authority.

What to watch in the final week

At this point, the most useful information will not come from counting every new poll. Four signals matter more.

First, watch whether the generic ballot changes materially from its current high-single-digit Democratic range. A stable number would suggest that the national judgment is largely formed; a sharp late change would require us to rethink the current picture.

Second, watch the Senate races for divergence. If Texas, Iowa, Alaska, Maine, Michigan and Ohio continue to move together despite very different candidates and political cultures, the national environment is doing more of the work. If they separate, candidate-level competency, reliability and local identity are reasserting themselves.

Third, watch the institutional disputes. The California ballot demands and Alaska certification lawsuit could remain narrow legal matters, or they could become broader signals that influence how voters think about election legitimacy. The more partisan actors frame administrative procedures as instruments of political power, the more difficult it becomes to preserve trust in those procedures independently of the result.

Finally, watch what the campaigns promise to do after November 3. The AP-NORC priorities data suggest an electorate that is less interested in partisan victory as an end in itself than in practical relief: costs, war, deficit reduction and some degree of bipartisan functioning. That creates a measurable post-election trust test. Whichever party gains power will have to decide whether to interpret victory as permission to intensify conflict or as an instruction to solve the problems voters have identified most clearly.

A week from now, the campaign will be almost over. The deeper experiment will just be beginning. The election will tell us who receives authority. What it will not tell us automatically is how much trust comes with it - or how quickly that trust can be lost if the winners misunderstand why they were chosen.

The final-week trust map

The election now contains four overlapping trust tests. None of these is a candidate score; each identifies the relationship currently under pressure.