Brief us

WP-30: Doctrine

Running as the institution

Eight in ten Americans say political institutions are run by wealthy and powerful interests. Six in ten still say hard problems need professional institutions to solve them. Campaigns that speak to both findings reach the voters who hold them.

By Brushworks research deskNovember 2026Working paper

Abstract. Distrust of institutions is near its recorded floor and is shared across both parties, yet majorities still say complex problems need professional institutions to solve them. We set the two findings side by side, read the 2026 primaries and the shrinking premium for moderation, name three errors candidates and institutions make in this climate, and describe a discipline built on checkable local results, an enlarged electorate and a sealed control.

In brief

The finding
The anti-establishment vote is not one bloc. Most of it believes the system is tilted and still wants it to work. That overlap decides low-turnout primaries when someone brings it into the room.
For a campaign
Do not run on standing and do not borrow the insurgent’s voice. Run on one checkable local result per household, answer the predictable charge before it lands, and enlarge the primary electorate.
For a company, an association or a public body
The same climate applies to a rate case, a siting or a ballot question. Lead with the local footprint and a disclosed name, not with national standing.

Anti-establishment sentiment is the climate of this decade, and it runs through both parties. It is also less uniform than it sounds. This paper separates the grievance from the program, shows where the candidate or institution with a record still has ground to stand on, and sets out the discipline that holds it.

The climate, measured

Average confidence across the fourteen institutions Gallup tracks stood at 27 percent in June 2026, one point above the lowest reading on record. Confidence in Congress was 9 percent.1

Confidence and trust in 2026: national institutions against local government. Basis: official, Gallup as published or as reported, field dates in notes 1 and 3.

The distrust is specific about its object. In a national poll at the end of 2025, 80 percent agreed that political institutions are controlled by wealthy, powerful interests, including 91 percent of Democrats and 75 percent of Republicans. Eighty-two percent agreed that elites do not understand everyday life. Only 26 percent agreed that institutions usually act in good faith.2

The same poll found something the first set of numbers tends to bury. Fifty-eight percent agreed that complex problems require professional, organized institutions, and 56 percent agreed that many policies are beyond what ordinary people have the information to judge.2 In Gallup’s September reading, 67 percent of adults trusted their local government, well above any national body.3

The grievance and the program, from the same national poll. The first three rows describe the complaint; the last two describe a public that still wants institutions to work. Basis: official, Economist/YouGov poll, December 26 to 29, 2025 (note 2).

A scholar of the subject defined populism two decades ago as a thin ideology: the pure people against a corrupt elite, with room left for almost any program.4 The numbers fit the definition. The grievance is near-universal. The program is not.

What the 2026 primaries showed

The pressure ran through both parties. Through the end of June, candidates endorsed by the main organizations of the Democratic left had won or advanced in 28 of 51 open contests, and three sitting members had lost their primaries to challengers from their left. On the Republican side, two sitting senators endorsed by their own party’s Senate committee lost to challengers backed by the president.5

Two details matter for anyone planning 2028. About 80 percent of those left-endorsed candidates ran in districts rated safe for their party, where the primary is the election and a small electorate decides it.5 And the general-election premium for moderation, long the center’s best argument, narrowed: one post-election analysis puts it at up to about three points in 2026, down from nearly five across 2018 to 2024, on preliminary counts.6 The older literature found a larger penalty for nominating from the edge.7, 8 The premium is real. It is no longer large enough to be the whole case.

The general-election premium for nominating a moderate, in percentage points: the 2018 to 2024 average against the 2026 preliminary estimate. Basis: record, one analyst’s published estimate, preliminary, with ballots outstanding (note 6).

Three errors

Running as the institution. “Experienced,” “responsible” and “endorsed by” are claims of standing. To a voter who believes the institution serves someone else, standing is the charge, not the defense.

Borrowing the register. A candidate with a record who adopts the cadence of an insurgent is caught at once. Voters forgive a great deal. They do not forgive being handled.

Arguing with the grievance. Eighty percent is not a misunderstanding to be corrected. Where the complaint is true, the campaign that says so first owns the next sentence.

The discipline

Find the pragmatic majority inside the anti-establishment vote. The two sets of poll numbers describe an overlap: people who think the system is tilted and still want it to work. They are distinguishable from the committed by behavior: what they read, whether they vote in primaries, whether a specific local result moves them more than a national argument. Research on the electorate’s middle finds most of these voters are moderate rather than inattentive or incoherent.9 They are also the voters least likely to be on anyone’s primary contact list.

Enlarge the electorate. A safe-seat primary is decided by a small and attentive electorate. The candidate with a record usually loses that room and wins the district. The work is to bring the district into the room: regular general-election voters who have never been asked to vote in a primary, reached early, about the thing that affects them.

Be checkable. “Trust me” has no market. “Check me” does. One result per household, with the date and the source: the bridge, the clinic hours, the rate that did not rise. This Review has argued the case for one sentence per voter; here the sentence is a receipt.

Be local. The trust that survives is close to home. An institution’s national standing is a liability in this climate and its local footprint is an asset. Lead with the branch, the plant, the payroll in the county.

Be early. The charge is predictable: bought, distant, part of the problem. The inoculation literature is sixty years old and has not been overturned: the attack answered before it lands is the one that does not.10 The answer includes the concession.

Contrast on delivery, from the public record. The honest contrast with an insurgent is not ideology. It is what has been delivered, what would be, and at whose cost. Every line sourced, nothing about the opponent that cannot be shown.

Measure. A tenth of every cohort is held out. If the receipt does not move the pragmatic voter against the control, the program changes the receipt, not the audience.

Where trust still lives

National distrust is not distributed evenly across the things a voter can see. The same public that gives Congress single digits gives its local government a majority, and the gap is not new. A long line of research on political trust finds that trust is a judgment about performance and proximity: people extend it to institutions whose work they can observe and withhold it from institutions they know only through the news.11

That finding is the practical center of this paper. An institution with a national reputation and a local footprint should run on the footprint. A candidate with a record should run on the part of the record a household has touched. The reason is not modesty. It is that the local result is the only claim in this climate that the voter can check without trusting the person making it.

What tends to move three groups inside an anti-establishment electorate, by kind of claim. Basis: illustrative, the shape the firm reads in its own programs, not a measured result; each program measures its own against a sealed control.

What counts as a local result. It is specific, recent and visible: the clinic that opened late hours, the bridge that reopened, the rate that was held, the plant that kept its second shift. It names a place and a date. It can be confirmed by someone the voter knows, which is why the messenger matters as much as the line. A result that has to be explained is not local enough.

What does not count. Endorsements, credentials and seniority are claims of standing, and the previous section explained why standing is the charge rather than the defense. National accomplishments count only where they arrive as a local fact.

How it is found. The program reads behavior to learn which result matters to which household: the commuter who crossed the bridge twice a day, the family that used the clinic, the employee at the plant. It then tests, against a held-out tenth of each cohort, whether hearing the result moves what that household believes. Results that do not move anyone are retired from the program even when they are true and impressive, because a receipt the voter does not care about is not a receipt.

The discipline is unglamorous. It is also the one approach in this environment that improves with scrutiny rather than suffering from it.

A worked example: a primary in a safe seat

What follows is a composite, with illustrative values and no candidate details, of the situation this paper describes.

An incumbent with a long record faces a primary challenge from an insurgent in a district the party always wins. The primary electorate in past cycles has been small, a fraction of the general-election electorate, and older and more attentive than the district as a whole. The challenger's coalition is energized and concentrated in that room.

The first decision. Win the room, or enlarge it. The terrain brief finds that the incumbent trails among habitual primary voters but leads comfortably among voters who vote in every general election and have never voted in a primary. The decision is to enlarge the electorate.

The program. Regular general-election voters who have never voted in a primary are identified from vote history and reached early, about the primary itself: that it decides the seat, when it is, how to vote in it. Persuasion for them is one checkable local result each, from the incumbent's record: the clinic, the road, the plant.

What is avoided. Claims of standing. Endorsements from the party establishment are not featured to these voters, because the program's measurement in the first weeks shows they do not move them.

What the control shows. Turnout among treated first-time primary voters rises against the held-out tenth, and support among them favors the incumbent. The room is larger than in any recent primary.

What it does not show. That the approach would work everywhere. In a district where the incumbent had no checkable record, there would have been nothing to say. The method depends on a record that survives scrutiny.

Common mistakes

Leading with endorsements. To a voter who distrusts institutions, an endorsement is evidence for the charge.

Borrowing the insurgent’s voice. Voters recognize a borrowed register immediately and do not forgive being handled.

Fighting in the room the challenger owns. A small, attentive primary electorate favors the challenger. Enlarging it is often the better path.

Running without a record to check. The method depends on results a voter can verify. Without them, the first task is to deliver some.

Implications

For a campaign. Decide before the first poll whether the race will be won in the room or by enlarging it. Staff and budget follow from that answer.

For a company, an association or a public body. The same climate applies to a rate case, a siting or a ballot question. The institution that leads with standing loses to the one that leads with a checkable local result and a disclosed name on the message.

For both. None of this is a way around the grievance. It works where there is a record to check. Where there is not, the first task is to build one.

Further reading

For readers who want to go further, these are the works this paper draws on most, or that argue the case at greater length. Each is listed with a line on why it is worth the time.

  • Hetherington, M. J. (2005). Why Trust Matters. Princeton University Press. Why trust in government matters for what it can accomplish.
  • Achen, C. H., and Bartels, L. M. (2016). Democracy for Realists. Princeton University Press. Why voters punish incumbents for conditions.
  • Tyler, T. R. (2006). Why People Obey the Law. Princeton University Press. How fair process earns acceptance of unwelcome outcomes.

Notes and references

  1. Gallup, confidence in institutions, June 1 to 15, 2026, 1,001 adults, as reported by Axios, July 14, 2026.
  2. YouGov, Economist/YouGov poll, December 26 to 29, 2025.
  3. Gallup, Governance survey, September 1 to 17, 2026.
  4. Mudde, C. (2004). The populist zeitgeist. Government and Opposition, 39(4), 541–563.
  5. FiftyPlusOne, “We tracked every candidate in the 2026 primaries,” July 31, 2026; data through the June 30 primaries.
  6. The Argument, “Do moderates still have an electability advantage?”, November 12, 2026; preliminary, with ballots outstanding.
  7. Hall, A. B. (2015). What happens when extremists win primaries? American Political Science Review, 109(1), 18–42.
  8. Canes-Wrone, B., Brady, D. W., & Cogan, J. F. (2002). Out of step, out of office: electoral accountability and House members’ voting. American Political Science Review, 96(1), 127–140.
  9. Fowler, A., Hill, S. J., Lewis, J. B., Tausanovitch, C., Vavreck, L., & Warshaw, C. (2023). Moderates. American Political Science Review, 117(2), 643–660.
  10. McGuire, W. J. (1964). Inducing resistance to persuasion: some contemporary approaches. Advances in Experimental Social Psychology, 1, 191–229. See also WP-04, Inoculation at scale.
  11. Hetherington, M. J. (2005). Why Trust Matters: Declining Political Trust and the Demise of American Liberalism. Princeton University Press.

Basis of figures. Polls: as published or as reported, with field dates in the notes. Primary results: one tracker’s count through June 30, 2026. The moderation premium: one analyst’s preliminary estimate. No Brushworks program figures appear in this paper.

Next paperThe Review

Start a conversation

Brief us.

Tell us the race, the calendar and the margin. Within a week we will tell you whether we can move it and what it would take. Candidate work and independent-expenditure work are taken by separate teams; say which you are. Governments and institutions: tell us the mandate and the jurisdiction, and we will tell you what registration it requires before we say anything else.

Send a briefbrief@brwrks.com

  • Candidate practice
  • Independent-expenditure practice
  • International practice

Brushworks Strategies Inc., a Wyoming corporation · Washington, New York, Ottawa, London, Munich, Dubai, Tokyo