The Supreme Court does not allow cameras, so on the morning of January 21, 2010, the country learned what happened inside Room One the way it always does: from reporters running down marble steps to a bank of pay phones and producers waiting outside with earpieces. Inside, Anthony Kennedy read a summary of an opinion that ran ninety-some pages including dissents, in a case that had started, two years earlier, over something almost quaint — a nonprofit called Citizens United wanted to make a ninety-minute documentary called "Hillary: The Movie" available on cable video-on-demand during the 2008 Democratic primary, and the Federal Election Commission said no, that counts as an electioneering communication within thirty days of an election, and BCRA bans corporate treasury money for that. A dispute about a made-for-DVD hit piece on Hillary Clinton became the vehicle for rewriting the campaign finance code of the United States.

What the Court actually did, on the record, is narrower than most people carry around in their heads. Five justices to four, Kennedy writing, struck Section 203 of the Bipartisan Campaign Reform Act and overturned Austin v. Michigan Chamber of Commerce (1990) along with part of McConnell v. FEC (2003), holding that corporations and unions have a First Amendment right to spend unlimited money from their general treasuries on independent political advertising. It did not touch the ban on corporations giving money directly to candidates — that stayed illegal. And on disclosure, the Court went eight to one the other way: only Clarence Thomas thought donors funding these ads should get to stay anonymous. The ruling that gets blamed for dark money mostly wasn't about dark money at all.

The gap between memory and record gets wider from there. The mechanism everyone means when they say "Citizens United" — the Super PAC, the vehicle that can take unlimited checks from a single donor and spend it praising or savaging a candidate — doesn't come from Kennedy's opinion. It comes from a D.C. Circuit case decided that March, SpeechNow.org v. FEC, which took Citizens United's logic about independent expenditures and used it to strike down the contribution limits on groups that only make independent expenditures. Citizens United built the argument. A different court, months later, built the object. The country has spent sixteen years blaming the blueprint for the building.

John Paul Stevens wrote the dissent, joined by every other Democratic appointee on the bench, and it is worth reading for the specific fear in it: not that corporations would openly buy candidates, which the ruling didn't permit, but that the sheer scale of available treasury money would let a handful of institutions drown out everyone else's version of political speech. Twenty-two words of prophecy versus ninety pages of doctrine, and the doctrine won five to four.

The case for the ruling, made honestly, doesn't rest on loving corporations. It rests on a fact reform advocates don't love to sit with: contribution limits and spending caps, whatever else they do, also entrench whoever already holds the seat. An incumbent has name recognition, a donor list, a franking privilege. A challenger has none of that and needed some way to compete with money the old rules made hard to raise. John McCain, who co-wrote the law the Court gutted, thought the answer was ceilings. The Court's majority thought the answer was doors. Both were arguing, in good faith, about how to keep politics from being purchased outright — they just disagreed on which purchase to worry about.

Here is the part the 2010 argument didn't anticipate, and it's the part that matters this month. The Super PAC apparatus that grew out of Citizens United and SpeechNow doesn't answer to a party chairman, a platform committee, or a primary electorate in the way old-fashioned hard money did. It answers to whoever wrote the check. That means a candidate can now build a financial operation entirely outside the institutional party — legally barred from "coordinating" with it, which in practice means separated by a firewall thin enough to read through — and use it to run past every gatekeeper the party used to have.

Run the line to today. A promise like sending voters five-thousand-dollar checks, the kind that would have died in a Ways and Means markup in 2005 without ever reaching a floor vote, survives instead as a live argument splitting the party's fiscal conservatives from its populists. The fiscal hawks control the committee rooms. They no longer control the money.

I want to be honest about how long this causal chain runs, because a columnist who pretends Kennedy's opinion in 2010 foresaw a specific dollar figure in 2026 is selling you something. The line from the ruling to the check is not straight. But the wiring is the same wiring: independent money, raised in whatever size a donor wants to write, spent on behalf of whoever it favors, with no party committee holding the other end of the wire. That is what changed in January 2010, whatever the headline said at the time.

The old system let party elders kill a bad idea by starving it of money before it reached a primary stage. The new one lets a bad idea, or a popular one, raise its own army and skip the elders entirely. Citizens United didn't invent populism inside the Republican Party. It just made sure that when populism showed up, it wouldn't need the party's permission to pay its own bills.

The photo that started the latest round is almost aggressively ordinary: a stand of University of Texas students in burnt orange, several of South Asian descent, cheering a fourth-quarter comeback against Ohio State. The Longhorns won 24-23 on Saturday. By Sunday morning, Bo French — the Republican nominee for Texas Railroad Commissioner — had turned that photo into a caption: proof, he wrote, of "how many Americans have been displaced by foreigners." He has since called the students in it, and the country they're citizens of, "the problem."

This was not French's first pass at the genre. He has called Native Americans "third world savages" who should be deported. He has proposed banning Muslim-country nationals from owning Texas property. He once ran a social media poll asking his followers whether Jews or Muslims posed the bigger threat to America — a poll that got him pushed out as Tarrant County GOP chair by Lt. Gov. Dan Patrick himself. He has called, separately, for the deportation of a hundred million people, a figure that necessarily includes American citizens, because there aren't a hundred million unauthorized immigrants in the country and French knows the math as well as anyone.

Give the hardliners their best version of the argument, because it deserves to be made honestly before it gets dismantled. Immigration enforcement in this country has been, for decades, a bipartisan performance of concern layered over administrative paralysis — a Congress that legislates nothing and an executive branch that governs by memo, in both directions. Voters who feel that strain in their wages, their schools, their emergency rooms are not wrong to feel it, and a political class that responds with euphemism has earned the contempt it gets. Blunt rhetoric, on this reading, is not bigotry but honesty with the varnish stripped off — the cost of forcing a conversation the establishment would rather manage than resolve. It is a real argument. It is also not the argument French is making.

Because the office French is running for regulates oil and gas production in Texas. It does not write immigration law, adjudicate asylum claims, or set enforcement priorities at the border. State Rep. Jared Patterson, no one's idea of a squish, made exactly this point in the same breath he called the post racist: "The Texas Railroad Commission regulates our oil and gas industry. It does not set federal immigration policy." Whatever French is selling, it isn't a policy program. A hundred million deportations is not a plan; it's an aesthetic, borrowed from a debate that his own office has no jurisdiction over, deployed against college students at a football game because the students were the nearest available prop.

The condemnations that followed were fast, specific, and came from Republicans with nothing obvious to gain. Senator John Cornyn called it what it was and asked, pointedly, whether other officials would look away "while it becomes normalized." House Speaker Dustin Burrows said you can oppose immigration policy without demeaning people over their ethnicity — a distinction French's campaign appears constitutionally unable to make. State Rep. Lacey Hull said she saw Texans in that photo, not a problem. Former Speaker Dade Phelan called him unfit for office.

Here is the part worth putting on the books, because it is the part that will still be true after the news cycle moves on. Dan Patrick demanded French's resignation from a county party post over the Jewish-Muslim poll. He has said nothing about French's nomination to statewide office over a much larger offense. Governor Abbott has said nothing at all. Attorney General Ken Paxton, who has actively campaigned for French, issued a statement condemning "racism in all its forms" without naming him — and used the same statement to take a swipe at a Democratic opponent instead. Cornyn asked the right question. The silence of the men who actually control the party's machinery is the answer.

French, for his part, did not walk anything back. He called Cornyn a RINO cheering for "foreign nationals over American citizens," and found defenders in former state party chair Matt Rinaldi and Rep. Andy Hopper, who called Cornyn a "smug globalist" — a senator who, they note, lost his own primary by twenty-eight points, which is meant as an insult and is actually the more interesting fact. French didn't win his primary on oil and gas expertise either. He beat an incumbent commissioner by running almost entirely on culture-war provocation, while a more substantive primary challenger that same cycle lost by thirty points and later joked, without much bitterness, that she'd have done better talking about the wall.

That is the structural story underneath the individual one. A party's stated principles mean nothing if the primary electorate rewards the candidate who abandons them and the leadership only finds its voice after the ballots are already printed. Texas Republicans said the right things this week, on the record, by name. The question the next primary will actually answer is whether saying them costs French anything at all.

On Thursday night, in a ballroom in New York, Barack Obama told a room of Democratic donors that artificial intelligence is "moving very fast in private hands" and that if the party doesn't get on top of it, the technology "can be dangerous." He told Hakeem Jeffries that if Democrats retake the House in November, the first job is a framework: a public conversation, built and led, not left to whoever gets there first. He told 2028 hopefuls to make AI a central plank, not a footnote. Nobody leaked this to embarrass him. The New York Times got a partial transcript from his own office, which tells you Obama wanted it read.

He is not wrong about the stakes, and he is not wrong that nobody in either party has done the work. But it is worth being precise about what he actually diagnosed, because the easy version of this story — Democrats are the adults, Republicans are the culture-war children — survives about as long as it takes to read the second paragraph of the Politico account of that same fundraiser.

There is a serious case for the Republican posture, and it deserves to be stated before it is dismissed. A party's electoral coalition is a fragile thing, built over cycles, and there is no evidence that swing voters in Erie or Pinellas County are waiting on an AI liability framework to decide how they vote. Culture and economic-security questions are cheap to communicate and expensive to abandon; AI policy is the reverse, hard to explain and harder to vote on, and genuinely unsettled even among the people who study it for a living. And Trump's framing — that the real contest is the race against China for who controls the technology, not the race to regulate it — is not merely evasion. It is an industrial-policy argument with real adherents, including plenty of Democrats who worked in Obama's own administration. Regulation that slows the frontier lab in San Francisco does not slow the one in Shenzhen. That argument deserves an answer, not a shrug.

Here is the answer. Treating "beat China" as the entire policy is like answering every domestic question in 1955 with "beat the Soviets": true as a stake, useless as a plan for what actually gets built, who is liable when it breaks, and what happens to the radiologist Obama mentioned, whose X-rays AI is already learning to read.

The public is not confused about the stakes, even if Washington is. A Fox News poll from April found 63 percent of voters believe AI harms privacy and 56 percent believe it harms job creation, against just 21 percent who think it helps. A July Fox poll found 70 percent oppose a data center being built in their own community. Those numbers do not sort by party. They are the numbers of a country that has noticed something large moving and wants to know who, if anyone, is watching it.

Democrats, for the record, do not have this solved either, and Obama's own remarks are an admission of that. Josh Shapiro, a governor who spent a year courting AI investment for Pennsylvania, is now posting on X that Washington needs "guardrails," while Bernie Sanders and Greg Casar are pushing a permanent ban on advanced AI development that most of the party's own donor class would treat as the extinction event for the industry rather than the risk it was meant to answer. The Klobuchar-Thune bill in the Senate, the one actual bipartisan attempt at a framework, is stuck on the two questions that matter most: whether AI companies bear legal liability for harm, and whether states get to regulate on their own if Congress won't. Nobody has answered either. That is not a Democratic advantage. It is a leadership vacuum Obama is trying to fill from outside government because nobody inside it has bothered.

The part that should worry Republicans has nothing to do with elections. The story of the administrative state's growth over the last century is not a story of bureaucrats seizing power in the dead of night. It is a story of Congress declining, cycle after cycle, to answer hard questions, and someone else answering them instead: an agency, a court, a governor, in this case whichever party actually shows up. If the House GOP treats AI the way it has treated entitlement arithmetic — as a subject too radioactive to touch honestly — the vacuum does not stay empty. It fills with executive orders, state patchworks, and consent decrees written by regulators nobody elected. The exact outcome conservatives have spent forty years warning about, arriving this time through their own absence rather than the other side's ambition.

Obama compared AI to nuclear power: a technology that transforms everything it touches and carries no natural ceiling on its own. Nuclear power got a framework because Congress in 1946 and again in 1954 decided the question was too important to leave to the labs that built the bomb. AI is getting no such thing from the party currently in the White House, and the working theory in Republican circles seems to be that dominance is a strategy and legislation is somebody else's problem. It is a fine theory right up until the first successor to Klobuchar-Thune passes with only one party's fingerprints on it, and Republicans discover they spent the decisive years arguing about 1954.

Dan McKee had been governor of Rhode Island for five years, had no scandal attached to his name, and had done the ordinary competent things governors do — balanced budgets, cut ribbons, avoided the papers. Then he became the first sitting governor in eight years to lose his own party's primary. Nobody indicted him. Nobody caught him in a lie. His own voters simply decided they were finished with him, which is a stranger and harder thing to explain than a scandal.

He was not alone. Nine House incumbents lost their party primaries this cycle — six Democrats, three Republicans, and by one count the tally has since climbed to ten. State legislatures saw 164 incumbents fall in the 43 states that have voted so far, a number that touched 40 states and both parties almost evenly: 124 Republican losses to 50 Democratic ones nationwide, once statewide races are folded in. County attorney races, the most local office on most ballots, produced a 62.5 percent incumbent loss rate — five of eight seats changed hands before a single general-election vote was cast.

The strongest response to all this is also the most boring, which is usually a sign it deserves to be taken seriously. Incumbency remains one of the most durable advantages in American politics; 96.4 percent of state legislative incumbents who sought re-election this year still won their primaries. The 2016 cycle saw 198 legislative incumbents lose — more than this year's 174 — and nobody wrote that year's obituary for representative government. Anti-incumbent sentiment tracks perceptions of bad times more than any fixed institutional rot, and those perceptions are famously colored by partisanship: a voter's sense that "things have gotten worse" depends heavily on whether her party holds power. Local scandals, local redistricting quirks, a bad map drawn by the other side's legislature — these explain most individual losses far better than any grand theory of a system in crisis.

Except 2026 was not a redistricting year. The maps that decided these primaries were the same maps that protected these same incumbents two years ago. That excuse, available in 2022 and usable again after 2030, sits idle this cycle. And the losses did not sort themselves by party the way a genuine wave election does — a punishment of the governing side for inflation, or the opposition for overreach. They ran through both parties in the same summer, at the state legislative level and the county attorney level and the U.S. House, which is not the signature of one electorate's grievance. It looks more like a shared appetite.

What the raw numbers miss is the behavior of the 96.4 percent who won. A 2018 study of congressional incumbents found that the credible threat of a primary challenge changes how members vote long before any challenger files paperwork — members "continually adapt to the possibility" of being primaried, adjusting positions and votes to close off the opening. Chuck Schumer, of all people, put the mechanism plainly: primaries are decided by "the vast majority of Americans" staying home and "the 10 percent at each of the two extremes" showing up, which means a member's real constituency on any contested vote is rarely the district. It is the sliver that turns out in June.

That is the finding worth sitting with, more than the loss tallies themselves. Congress does not need a wave of incumbents actually defeated to be reshaped by anti-incumbent sentiment. It needs incumbents who believe they could be. Fear is cheaper than defeat and works just as well on behavior — a legislature full of members quietly recalculating every appropriations vote against a primary electorate that skews to the loudest ten percent on each side is a legislature that has already outsourced its judgment before the roll is called.

I grew up watching a congressman from my part of Ohio survive one bad primary scare and never cast an uncomfortable vote again — not on entitlement arithmetic, not on a base closure, not on anything his district's most engaged voters had opinions about, whether or not those opinions matched the district's interests. He kept his seat for another decade. Nobody ever had to name what changed. That is the part a loss column can't capture: the votes never taken, the amendments never offered, the retirement announced instead of the fight picked.

None of this requires apocalypse. The republic survived 198 primary losses in 2016 and will survive 174 more this year. But an institution built on the assumption that representatives answer to their whole constituency does not need to be destroyed to stop functioning as designed — it only needs enough of its members correctly calculating that the whole constituency isn't who's watching. McKee lost an election. The other 96.4 percent may simply have learned the lesson he taught them for free.

Denny Hoskins spent Thursday morning explaining himself to seven judges who did not appear to be in a forgiving mood. Missouri's secretary of state had been ordered to appear before the state Supreme Court and account for why he should not be held in contempt, after he kept pushing the Republican-drawn congressional map even after the court told him, unanimously, to stop. Hoskins had posted on X that he looked forward to his day in court. He got it.

The ten days before that hearing are worth laying out in order, because the order is the story. On September 3, the Missouri Supreme Court ruled unanimously that the new map — HB1, drawn to convert one of Emanuel Cleaver's Democratic seats into a Republican one — could not be used, because a referendum petition against it was "legal, sufficient, and timely" under the state constitution, and no act of the general assembly is exempt from that citizen check. The old 2022 map, six Republican seats to two Democratic, snapped back into place.

Two days later Justice Brett Kavanaugh declined to disturb that ruling on the Supreme Court's shadow docket. Minutes after that, a federal district court issued a temporary restraining order requiring the state to use the new map anyway, on the theory that running a primary under one map and a general election under another would violate voters' equal protection and Article I rights. On September 9 the Eighth Circuit declined to stay that order. Missouri now has two federal-adjacent rulings pointing in opposite directions and a state official under contempt review for picking a side.

The strongest case for Hoskins is not the one he made in his X posts, and it deserves to be stated properly. The Constitution's elections clause gives "the Legislature" of each state the power to prescribe the times, places, and manner of congressional elections — and a citizen referendum petition, however lawful under Missouri's own constitution, is not the legislature sitting in session. There is a genuine harm underneath the abstraction, too: thousands of Missourians voted in an August primary drawn under HB1's lines, and if the November general reverts to the 2022 map, some of them will find themselves choosing a different representative than the one their primary ballot was building toward. That is not a hypothetical inconvenience. It is the exact injury the federal district court cited when it ordered the new map used.

The elections clause argument runs into a wall the U.S. Supreme Court already built. In Arizona State Legislature v. Arizona Independent Redistricting Commission, the Court held that "the Legislature" in the elections clause means whatever lawmaking process a state's own constitution establishes — and that a state constitution may reserve part of that power to the voters directly, through initiative or referendum. Missouri's constitution has done exactly that for more than a century. A referendum petition is not an outside interference with the legislature's power. Under Missouri law it is the legislature's power, exercised by the only body the state constitution says gets the final word.

The primary-versus-general problem is harder, and I will not pretend the district court's discomfort with it is manufactured. Splitting the electoral calendar across two different maps is genuinely disruptive, and I am not certain any court, state or federal, has a clean way to unwind it once a primary has already been run. But notice whose decisions produced that mess. The general assembly passed HB1 fully aware a referendum clock was running against it, then let the primary go forward on the new lines rather than waiting out the challenge, betting that litigation would move too slowly to matter. Hoskins then kept certifying the new map after the state's highest court told him not to. The collision between the primary and the general is not two constitutional commands meeting by accident. It is what happens when an elected official tries to outrun a court order instead of complying with it.

The part the shadow docket fights obscure is this: whichever map governs Missouri's House races this November, the referendum itself is still on that same November ballot. The Missouri Supreme Court's ruling didn't just block HB1 for this cycle — it put the underlying question of which map Missouri should use permanently in front of the same voters who will be casting ballots in the districts those maps create. Every filing, every stay application, every contempt hearing is table-setting for a vote that was always going to happen regardless of who wins the interim skirmish.

That is the design, not the failure. Judges can pause a map and legislatures can gerrymander one, but Missouri's constitution built a valve neither branch controls: the citizen referendum, reserved by the people to the people, running on its own separate clock. The courts have spent a week and a half fighting over which map decides August's and November's districts. They have no say over the ballot question underneath it. Whatever the Eighth Circuit does next, whatever the Supreme Court eventually decides to take up, that vote happens on schedule — and it belongs to Missourians, not to seven judges in Jefferson City or nine more in Washington.

On September 2, a man named Steve Dakopolos stood on the Capitol steps in Jefferson City holding a stack of petition signatures, one voter among the thousands who spent a year collecting names to force a statewide vote on Missouri's new congressional map. Eight days later the Missouri Supreme Court gave him what he wanted, blocking the GOP-drawn districts from use in November and ordering the question onto the ballot. On roughly the same set of facts, a federal judge let the state keep using the map anyway. Missouri now has one congressional map, two courts, and no agreement between them about which one governs the election that starts to matter in about five months.

The map's defenders have the better opening argument than most of their critics will admit. Missouri lawmakers enacted the new districts in a proper special session in 2025, following procedures the state's own high court had already upheld; it rejected, earlier in this same fight, claims that the session itself was illegal and that mid-decade redistricting is barred by the state constitution. The group Put Missouri First argues the result is "more compact and split fewer counties and cities than the prior map," and calls the barrage of lawsuits what any losing party is entitled to call it: politically motivated. Attorney General Catherine Hanaway went further, calling the court's final block "a full-fledged constitutional crisis." State Rep. Brattin, who helped found the Missouri Freedom Caucus, said voters shifted into new districts on short notice are being disenfranchised by judges rather than served by them. These are not frivolous complaints. A legislature that follows the rules and then watches its work vetoed by a court is entitled to ask what the rules were for.

But "more compact" is a claim you can check against a map, and the map does not cooperate. The ACLU's suit on behalf of Kansas City voters describes district lines that run through the city along what it calls racial lines: in one dense neighborhood, the new boundary places Black and white residents on opposite sides of the same railroad tracks. That is not the ordinary untidiness of drawing 435 districts across a continent. That is a specific, locatable seam, and it happens to run through the district held by one of Missouri's two Black members of Congress.

The process errors are worse than the politics. The same lawsuit alleges the legislature, in its haste to deliver the map Washington had asked for, double-counted certain residents, violating the basic requirement that districts carry equal population, and misallocated others so badly that one district is not even contiguous. Two disconnected pieces of land, one congressional seat. Whatever you think of the politics behind this map, a legislature that cannot draw contiguous districts has not been slandered by a court that noticed.

What makes Missouri instructive rather than merely embarrassing is the structure underneath the confusion, because there is one. The state Supreme Court's ruling rests on the state constitution, its compactness clause and its once-a-decade language, and under a doctrine lawyers call adequate and independent state grounds, that puts the ruling almost entirely out of the U.S. Supreme Court's reach. The federal judge, meanwhile, is working a different case entirely, one built on the U.S. Constitution's elections clause, which the Supreme Court already declined to read as barring state courts from policing their own redistricting maps, in a 2023 case out of North Carolina.

Two courts are not fighting over the same question. They are each answering the question assigned to them, and Missouri voters are the ones standing where the answers overlap.

That is not chaos so much as it is the federalism the founders actually built, running at full speed and full expense in a single state at once. A system with two sovereigns and two court systems will occasionally produce two rulings on the same map in the same week, and the fact that this looks like disorder to the people living through it does not make it a malfunction. It makes it Tuesday.

What should trouble both sides is a fact neither party's press release mentions: the Missouri Supreme Court had already rebuffed the map's opponents three separate times, on the special session, on mid-decade timing, and on an earlier compactness claim, before it finally found a violation serious enough to stop the ballots. A court that ruled for the legislature three times running is not the runaway partisan tribunal Hanaway describes. It is a court that kept giving the state's answer the benefit of the doubt until the answer stopped holding up.

I grew up in a state that has run its own version of this fight for twenty years, and I have learned to distrust anyone who tells you redistricting litigation is either a righteous crusade or an unelected coup, because it is usually a legislature that moved fast, a court doing the job the legislature left for it, and a lot of voters who will not know their own district until close to the filing deadline. Congress has the constitutional authority to set a national floor on how states draw these lines and has mostly declined to use it, which is how you end up with fifty different answers to the same question and, in Missouri's case, two courts inside one state giving different ones in the same month.

Hanaway calls the result a constitutional crisis. It is closer to the constitution working exactly as designed, slowly, expensively, and in public, which is not the same thing as working well. The actual crisis is smaller and more damning: five months before an election, nobody in Missouri, least of all the people who have to vote in it, can say with certainty which district they live in.

It was Easter weekend in 2007 when a Barton County patrol car pulled into Joe and his wife's driveway in rural Kansas. Their grandchildren were dyeing eggs. The couple owed $4,561 from an emergency hysterectomy three years earlier, they had missed a court date on the collection suit, and a judge had signed a warrant. They spent two days in jail before their son could scrape together bond.

The doctor who sued them was Roger Marshall, now the junior senator from Kansas. According to a New York Times analysis of state court records, Marshall filed more than 700 lawsuits against patients over unpaid bills during his career as an OB-GYN in Great Bend, charging 18 percent annual interest on the balances. Eighty-one of those suits ended in an arrest, most for missing a court date on the debt itself. Wages were garnished. Bank accounts were seized. This was not a fringe habit of one hard winter; it ran with the same regularity as the billing office, across decades.

The case for leaving this alone is not stupid, and it deserves to be stated at full strength rather than waved off. Marshall was a business owner before he was a senator, signing checks for a staff that grew from five employees to more than three hundred. Unpaid medical bills are a real cost, and a practice that eats them indefinitely does not stay a practice for long. Collection lawsuits over unpaid debt are legal in all fifty states; nothing in the Times' reporting suggests Marshall's firm broke the law that governed it. A senator who has met a payroll and chased a receivable, the argument goes, understands the economics of rural medicine in a way that career legislators do not, and Congress could use more members who have signed the front of a check instead of only the back.

Fair enough. But the defense collapses the moment you separate two things Marshall's advocates keep bundled together: charging interest on unpaid bills, and using the criminal justice system as a debt-collection tool. The first is ordinary business practice, however unwelcome to the patient. The second is a different instrument, and the distinction matters precisely because it is so easy to blur — a missed hearing on a $4,561 medical bill produces the same police car in the driveway as a missed hearing on a felony charge. Eighty-one times, in Marshall's court records, it did.

Then there is the matter of Marshall's own testimony against himself. In July, promoting his Credit Card Competition Act on Fox News Radio, Senator Marshall described interest rates of 20 to 30 percent charged by banks and credit card companies as "predatory loaning" — his words — adding, "that's something that the mob rates would offer as well." He was speaking about a $7,000 average household balance carried by struggling families. He was not speaking, apparently, about the 18 percent he charged patients recovering from hysterectomies and cardiac surgery, which falls just shy of the range he now calls loan-sharking when Visa does it. How many of the 700 patients he sued would have qualified, by the senator's own current definition, as victims of a predatory rate? The number is not in the court records, but the arithmetic is not hard.

So let the strongest version stand and answer it on its own terms. Business acumen is a legitimate qualification for legislating healthcare economics, and nobody should want a Senate composed only of people who have never sent a bill. What disqualifies is not the ledger. It is the gap between the standard Marshall enforced with a badge and the standard he now demands of Visa with a press release — a gap that is either hypocrisy or a genuine, unremarked change of conscience, and the senator has offered no public account of which.

Congress does not currently require a member who once ran a debt-collection operation against his own patients to disclose that fact when he votes on medical billing reform, hospital price transparency, or the collections practices the Consumer Financial Protection Bureau is authorized to police. There is no recusal norm, no affidavit, no line on a financial disclosure form that says "I sent 81 patients to jail; adjust your reading of my healthcare votes accordingly."

That is a gap in the machine, not a smear against the man who exposed it by running through it. None of this requires imagining a private financial motive behind any specific vote Marshall casts today; the record does not show one, and inventing it would be its own kind of malpractice. What the record shows is a senator whose past practice and present rhetoric on interest rates cannot both be sincere, and a Senate with no mechanism for making him choose. Put that on the books. Voters in Kansas can decide what to do with it this fall — but the disclosure should not have waited on a reporter to find the court file first.

The Missouri Supreme Court did not run for Congress, campaign for a seat, or submit itself to a single contested voter. Last week it paused the state's newly drawn congressional map anyway — the one the legislature passed this year to convert a competitive district into a safe Republican one. If the pause holds, Missouri could send one more Democrat to the House in 2026 than the legislature intended. Seven judges, elected under Missouri's retention system rather than chosen in a contested race, just moved a number that will show up in the Speaker's vote count.

This is not an isolated curiosity. Texas redrew its map for more Republican seats, California countered under Governor Newsom with a mid-decade map of its own, and Nevada Republicans have been maneuvering against Newsom's push in turn. North Carolina, meanwhile, is sending out mail-in ballots this cycle while litigation over the rules for counting them remains unresolved. None of these fights are happening in Congress. All of them will determine who sits in it.

The case against letting state courts referee any of this is stronger than its critics usually bother to make, and it deserves a fair hearing before it gets an answer. The Constitution's Elections Clause gives state legislatures, not state courts, the authority to prescribe the "times, places and manner" of federal elections. Justice Alito, joined by Justices Thomas and Gorsuch, pressed exactly this point in the 2020 Pennsylvania mail-ballot dispute: if a state court can override a legislature's election rules by invoking some open-ended state constitutional phrase — Pennsylvania's is "free and equal," Missouri's has its own — then the legislature's textual grant of authority is not worth the paper it is printed on. A rule that can be rewritten by seven judges any time they find it inconvenient is not a rule. It is a suggestion with a robe on.

There is a subtler version of the objection worth taking seriously, and it is not about legislatures at all. It is about accountability. The worry is that unelected judges are overriding the choices of elected representatives, substituting their preferences for the voters'. But Missouri's justices are not life-tenured. They face retention votes, as did the justices who paused Wisconsin's maps, and North Carolina's, and Ohio's, in earlier cycles. The honest version of the complaint is not that these judges answer to no one. It is that they answer to a different electorate, on a different clock, and that a legislature dissatisfied with the answer has no clean way to overrule it. That is a real institutional friction, and it is not the tyranny-of-the-unelected story its loudest advocates prefer.

Which makes the precedent do the real work. In Moore v. Harper in 2023, the Court rejected the independent state legislature theory outright, holding that state courts retain their ordinary power to review election laws against state constitutions; federal courts may step in only when a state court has plainly exceeded the "ordinary bounds of judicial review," a standard the Court left deliberately narrow. That single holding is why Missouri's pause is standing at all. Had the theory won, a state supreme court would have had no authority to touch a gerrymandered map no matter what the state constitution said about it.

Then set Rucho beside it. In Rucho v. Common Cause in 2019, the same Court held that federal courts have no power to police partisan gerrymandering at all; the question, it said, is political, beyond judicial management under the federal Constitution. Put the two rulings side by side and the arithmetic is not subtle. Rucho closed the federal courthouse to partisan gerrymandering claims. Moore kept the state courthouse open, on state constitutional grounds Washington cannot touch. If a gerrymander is going to be checked by anyone wearing a robe, it will be a state judge, applying a state constitution, because the Supreme Court itself has ruled out every other option.

North Carolina shows what this looks like when it is not merely about maps. Ballots are going out to mail voters this cycle while litigation over the rules for counting them remains unresolved — a live rerun of the dispute Alito flagged in 2020, when the Pennsylvania Supreme Court extended a mail-ballot deadline by invoking its state constitution over the legislature's own text. Whatever a court eventually decides about which ballots count and when, the deciding will happen in a state courtroom, under a state constitution, well after the ballots are already in the mail. Voters are participating in a process whose legal foundation is still being argued.

None of this is confined to courtrooms. Nevada's Republicans are running their countermeasure against Newsom's map through the legislative process rather than the judicial one, a reminder that redistricting is now fought on every available front. The courts are simply the forum where the fight has the most leverage this year, because Rucho left the federal door bolted shut.

Here is what should trouble a reader who cares about process more than outcome, whichever party a given map happens to favor. Congress has the power, under the Elections Clause itself, to set uniform federal standards for congressional redistricting: nonpartisan commissions, fixed criteria, a single national rule instead of fifty state answers arrived at by fifty differently selected judiciaries, some elected, some appointed, some retained for eight years and some for life. It has never used that power in any durable way. So we get Missouri's justices, standing for retention next, deciding a House seat that Congress could have taken off the table by statute and chose not to.

That is not a judiciary run amok. It is a legislature that would rather let judges take the blame for an outcome it lacked the will to legislate. Put it on the books: the quarrel is not really between courts and legislatures at all. It is between the Congress we have and the Congress that has spent three decades declining to be one.

The photograph is dated August 17, 2020, and it shows nothing more dramatic than a row of blue mailboxes outside a post office in Washington, sitting in the kind of heat that makes the asphalt shimmer. By that week the mailbox had become a partisan object. The new postmaster general, a Trump megadonor with no postal background, had spent the summer pulling high-speed sorting machines out of processing plants and hauling collection boxes off street corners in a cost-cutting drive that happened, by coincidence the administration insisted was total, to land three months before an election in which tens of millions of Americans were about to vote by mail for the first time because a pandemic had made the ballot box a health hazard.

More than twenty states sued. A federal judge in Washington state called the changes an attempt "to impede the operations of the USPS" and ordered them reversed. Under that pressure and a congressional hearing that did not go well for him, the postmaster general announced in late August that he would suspend further changes until after November. The mail got slower that fall, and it got there.

No court ever found that the sorting-machine purge threw the election, and no one seriously argues it now. The memory that survives is louder than the record: for a lot of people 2020 is remembered as the year Trump tried to steal the election through the mail, when what actually happened is narrower and more useful — a president tried to reach election administration through an agency he did not control outright, and the courts and the states pushed him back before it mattered. That is worth having on the desk now, because the sequel started in March, and it is not using the same tool.

This time there was no cost-cutting pretext. Trump's March executive order tried to build a centralized federal list of adult citizens and hand the Postal Service the authority to refuse delivery of a mail ballot to anyone not on it — a direct claim that the executive branch could set the terms of who gets to vote by mail, nationwide, by rule. In May a federal judge declined to block it, reasoning that USPS hadn't yet written the rule so the fight was premature. The very next day, USPS published the rule. By early June the administration was building the citizenship-verification list the order called for. On June 25, U.S. District Judge Indira Talwani read a thirty-seven-page ruling that took the shorter path: the president does not have the constitutional authority to regulate state elections this way, full stop, and she blocked the rule's key pillars through the midterms.

The administration's case for the rule, stripped of the politics around it, is not crazy on its face. Mail voting scaled up enormously and fast, first in 2020 and again since; a genuinely uniform, federally verified voter list would close gaps that fifty different state systems leave open by design, and DOJ's court filings insist election officials can absorb the change without the disruption the plaintiff states describe — that the states are, in the government's word, overstating it. If your priority is a single national standard for who can cast a mail ballot, routing it through the one federal agency that already touches every ballot in the country has a certain engineering logic.

The Constitution's answer to that logic was written before there was a Postal Service to argue about. Article I gives the states the power to set the times, places, and manner of elections, with Congress — not the president, and certainly not an agency head — holding the only federal override, and only for congressional races. An executive order cannot conscript the mail system into enforcing a voter list the states never adopted and Congress never passed. Talwani's ruling is not really about postal logistics. It's about which branch gets to touch that clause at all.

Here is where the sequel gets more interesting than the original. DOJ didn't simply appeal Talwani's ruling on the merits — it went to the 1st Circuit asking that her injunction be narrowed, citing the Supreme Court's 2026 decision in Trump v. CASA to argue that the twenty-four states and the District that sued, led by Minnesota's attorney general and joined by California, Massachusetts, Nevada, Washington, and eighteen others including North Carolina, can only win protection for themselves. Not the country. Themselves. If the 1st Circuit buys it, the same USPS rule that a federal judge called unconstitutional could take effect anywhere a state didn't happen to be a plaintiff, while staying dead in the states that sued in time.

That would produce something the 2020 fight never quite managed: a mail-ballot regime that depends on which side of a state line the envelope is dropped in, adjudicated not by what the Elections Clause says but by who filed first. North Carolina, one of the plaintiff states, is mailing general-election ballots this month regardless of how any of this resolves. The trucks leave the sorting center in Raleigh on schedule, same as they have for a hundred years, while lawyers in Boston argue over whether the rule governing what's inside those envelopes applies nationwide, applies to North Carolina specifically, or applies nowhere at all until October.

Five years ago the argument was whether an agency could slow the mail enough to matter. Now it's whether a single court can protect an election that crosses fifty jurisdictions, or only the ones smart enough to sue. The costume changed — sorting machines then, a citizenship list now — but the shape underneath is the one the framers built the Elections Clause to prevent: an executive branch reaching for a lever the Constitution left in the states' hands, and the states reaching back to take it away, case by case, mailbox by mailbox.

Steve Dakopolos stood outside the Missouri Capitol on September 2 holding a hand-lettered sign, one face in a slow crowd that had been gathering outside that building since March. He had collected signatures in the spring alongside more than 300,000 other Missourians who wanted a vote on the state's new congressional map. Three days later, the Missouri Supreme Court told him he was going to get one.

The court's ruling, unanimous, blocked Republicans' new gerrymander (House Bill 1) from use in the November 2026 midterms, and ordered the 2022 congressional map to stand until voters decide HB 1's fate at the ballot box. The justices held that the referendum power in Article III, Section 49 of the Missouri Constitution functions the way the Governor's veto does elsewhere in that document: essential, not optional, and not something a Secretary of State gets to wave away because the timing is inconvenient.

Missouri Attorney General Catherine Hanaway did not receive that reasoning kindly. "Never before in American history has a court overturned a congressional map after a primary and before a general election," she said, promising an appeal to the U.S. Supreme Court and calling the ruling a "full-fledged constitutional crisis." Her solicitor general, Louis Capozzi, invoked the Purcell principle before the justices, the doctrine that courts should not scramble election rules on the eve of a vote, lest they confuse voters and depress turnout. That is a real doctrine, protecting a real interest, and it deserves better than dismissal. A state does not want its congressional lines decided by injunction eight weeks before ballots print.

But that framing depends on a premise the court's own opinion refuses to grant: that HB 1 was ever the law to begin with.

Under the court's reading, and under its own precedent in Maggard, a referendum petition that is lawful, sufficient, and timely suspends the statute it targets the moment it's certified. That happened here as of December 9, 2025, months before a single ballot for 2026 was printed. HB 1 never took effect. The 2022 map was never displaced. There is no map being "overturned" close to an election, because the map in question was never operative in the first place. What actually sat close to the election, dangerously so, was Secretary of State Denny Hoskins's decision, made last month after sitting on the signatures for weeks, to declare the referendum unconstitutional and try to force HB 1 into use anyway. If there is a Purcell problem in Missouri, it was manufactured by the official who delayed the ruling as long as institutionally possible and then complained about the clock he'd run down himself.

The textual case for the court's holding is stronger than Hanaway's brief lets on, too. Missouri's constitution contains explicit carve-outs exempting certain redistricting plans, those drawn by an independent bipartisan citizens commission, from the referendum power. The legislature knew how to write an exemption when it wanted one. It didn't write one for maps drawn by itself. A court reading that omission as deliberate isn't inventing a check on the legislature; it's declining to invent an immunity the legislature never asked the voters to grant.

Put the politics on the ledger honestly, because pretending they aren't there insults the reader. Reverting to the 2022 lines likely costs Republicans a seat, probably the one held by Kansas City's long-serving Democratic congressman, whom HB 1 was drawn to dislodge. Democracy Docket's reporting notes the map was produced at President Trump's request, a national fight over the House majority imported wholesale into a state legislature's pen, exactly the kind of maneuver that turns a state's own institutional design into collateral damage in someone else's contest. Federalism exists to let states run their elections by their own rules. It does not survive a governing party treating those rules as an obstacle to route around whenever a national majority is on the line, of either party, in any state, at any time.

None of this settles whether gerrymandering itself is wise policy, or whether the 2022 map is fairer than HB 1. Reasonable people can and do argue both. What the Missouri Supreme Court actually decided is narrower, and I'd argue more defensible: that a state's constitutional veto mechanism survives contact with a legislature that dislikes the outcome, and that an executive officer cannot starve that mechanism by delay and then invoke the emergency his own delay created.

Ask three plain questions and check the answers yourself. Did the referendum gather enough signatures, over 300,000, well before Hoskins acted? It did. Did Hoskins wait until last month to reject it, months after certification was possible? He did. Did the court order anything other than what Missouri's own constitution says happens automatically once a referendum is lawful, sufficient, and timely? It did not.

Voters get their ballot question on November 3. Whatever they decide about HB 1, they will have decided it, and not a secretary who sat on their signatures until the calendar became his excuse.

Steve Dakopolos was still out there on September 2, standing in front of the Capitol in Jefferson City with a petition sign. It had been a year almost to the week since the rotunda first filled with protesters over House Bill 1, the congressional map Missouri Republicans muscled through in 2025 to redraw Kansas City's longtime congressman out of a district. The next day, the Missouri Supreme Court ended his year the way petitions almost never end. It won.

The ruling, handed down September 3 and signed by all seven justices, was not close on the law even if it will be loud in the politics. Article III, Section 49 of the Missouri Constitution authorizes a referendum on "any act of the general assembly," Judge Ginger Gooch wrote, and "no exception applies," which made the petition from People Not Politicians Missouri "legal, sufficient, and timely." A referendum-challenged law does not take effect until voters approve it, so HB 1 never took effect at all. The state reverts to the congressional lines drawn after the 2022 census. Missourians will decide the new map's fate on the November 3 ballot, the first time a congressional plan has gone before them since 1922.

Take the other side seriously before you take the ruling's, because the other side has real cards. Cole County Circuit Judge Daniel Green had ruled the opposite way below, finding that the U.S. Constitution's Elections Clause hands congressional redistricting to state legislatures alone, full stop. Attorney General Catherine Hanaway isn't bluffing when she calls this uncharted water. Missouri already ran its primaries under HB 1's lines, and she's right that no American court has overturned a congressional map after a primary and before the general. She's headed to the U.S. Supreme Court calling it a constitutional crisis. A voter who showed up in August to vote in one district and gets reassigned in November has a legitimate grievance, and "the court said so" is not an answer to it, only a citation.

There is a second argument underneath the loud one, and it deserves the label conservatives usually reserve for their own instincts: stability matters. Electoral maps that flip mid-cycle on judicial say-so are maps nobody can plan a campaign or a candidacy around. If courts can rewrite district lines nine weeks before a general election, every losing legislature has an incentive to try its luck in front of a friendlier bench next time. That is not a frivolous worry. It is the same worry the right has voiced for two decades about judges finding rights nobody wrote down.

But look at what the Missouri court actually did with the text, because it did not invent a check. The same constitutional article that gives the legislature the redistricting pen carves out an explicit exemption for maps drawn by an independent bipartisan commission, shielding those specifically from referendum. The framers of that provision plainly knew how to protect a redistricting plan from a public vote when they wanted to. They did not protect the legislature's own maps. A court that enforces the exception the voters wrote, rather than the one the legislature wishes had been written, is not overreaching. It is reading.

And the "chaos" now unfolding traces back further than Thursday's opinion. The petition was filed, certified, and sitting on Secretary of State Denny Hoskins's desk well before the primary. It was Hoskins who rejected it, and the court that corrected him. If there is a villain in the timeline, it is an executive officer who blocked a lawful referendum, not a judiciary that eventually noticed.

Here is where the case gets interesting for readers who don't much care about Missouri redistricting law, because it scrambles both sides' comfortable stories. Progressives who have spent a decade asking courts to be the permanent backstop against gerrymandering should notice that this court did the opposite of activism. It declined to draw a fairer map itself and handed the question to the one body above the legislature in a republic, the electorate. Conservatives who treat "the legislature drew it, so it stands" as a complete sentence should reckon with the fact that Missouri's own constitution, ratified by Missouri's own voters, built a referendum valve into the legislative process on purpose, the way it built in the governor's veto. An unchecked legislative power over its own district lines was never the deal Missourians signed.

None of this makes a referendum a cure. Ballot campaigns get bought by the same money that buys legislatures, ballot language gets litigated into incomprehensibility, and a "no" vote in November settles nothing about whether the 2022 lines are themselves any fairer. They were drawn by the same legislature, after all, just a friendlier one to sitting incumbents. Direct democracy is not a machine that produces good maps. It is a machine that makes bad ones answer to somebody.

Which is about the most a republic can honestly promise. On November 3, the district lines that stand or fall will be the ones Missouri voters put there, not a Cole County judge's reading of Article I, not a secretary of state's desk drawer, not a legislature that assumed nobody was allowed to check its work. Until then, the 2022 map governs, and the congressman Republicans tried to draw out of a job runs in the district he has held since 2005.

The photograph that ran with the story shows a man named Steve Dakopolos standing in front of the Missouri Capitol on September 2nd, holding a sign, in the particular late-summer light that makes every statehouse protest look like it has been going on forever. By then it nearly had. Missouri Republicans passed a new congressional map last year at President Trump's request. Voters spent the better part of a year trying to get a public vote on whether to keep it. Six days before the state's deadline to lock the November ballot, the Missouri Supreme Court told them they could.

The mechanics matter more than the headline. House Bill 1 redrew Missouri's congressional lines for the second time this decade, a move designed to convert a competitive Kansas City-area seat into a safe Republican one. Citizens gathered enough signatures to force a referendum — the direct, up-or-down vote Missouri's constitution reserves for the people when they distrust what the legislature has done. Secretary of State Denny Hoskins refused to certify it. A Cole County judge, Daniel Green, sided with him, ruling in late August that congressional redistricting isn't the kind of legislative act the referendum power reaches, and that swapping maps this close to an election would risk invalidating an August 4th primary already run. The Supreme Court heard the appeal on September 2nd and reversed him. HB1 never took effect. The 2022 map governs November, and voters get their say on the new one anyway, on the same ballot.

Here is the case for Judge Green's reading, stated as fairly as I can manage, because the people making it were not cranks. Redistricting is textbook legislative business — the General Assembly draws lines, the governor signs them, the process is done in the ordinary way lawmaking is done everywhere. Forcing a second vote on top of a completed legislative act, mid-cycle, six days before a ballot deadline, is not democratic hygiene. It is the kind of administrative chaos that produces exactly what Green warned about: a primary held under one map, a general election potentially held under another, thousands of already-cast primary votes sitting in a legal cloud nobody can fully clear. That is a real cost. Anyone who tells you courts can rearrange electoral maps in September without friction is selling you something.

There is a second inconvenient fact the gerrymander's critics don't advertise as loudly as they should. This was not HB1's first trip to court. Black voters in Kansas City sued separately, arguing the new lines violated the state constitution's requirement that districts be "as compact as possible" and that Missouri has no business redistricting twice in one decade at all. The Court heard that case in May and ruled against them — the map, on the merits, was lawful. So whatever the Court decided on September 2nd, it did not decide that HB1 is a racial gerrymander or a constitutional abomination. It decided something narrower and, I'd argue, more important: that a legislature cannot pass a law and then argue its way out of the one check the constitution puts directly in the voters' hands.

That distinction is the whole case. The Missouri Constitution doesn't leave "is congressional redistricting a legislative act" to inference. The document specifically exempts maps drawn by an independent bipartisan commission from the referendum power — meaning the drafters knew how to carve out an exception when they wanted one, and did not carve one out for maps the General Assembly draws itself. The Court called the referendum and initiative power an essential part of Missouri's lawmaking process, on par with the governor's veto. A governor's signature and a citizen's referendum are the same category of check, applied at different ends of the process, and no one seriously argues a legislature can pass a bill and then declare it immune from a veto because vetoes are inconvenient in September.

That is the answer to the timing argument too, and it is a harder answer than "the ends justify it." The chaos Judge Green worried about was manufactured, not discovered — the legislature could have passed HB1 with enough lead time to survive a referendum challenge and chose not to, and the Secretary of State compounded the delay by refusing certification for months before the courts got involved.

A constitutional right that evaporates whenever the other side runs the clock down is not a right. It is a formality.

None of this means the new map is bad policy, or that the old one is good policy, or that Missouri's Kansas City district should look one way rather than another. Those are questions for voters, which is precisely the point: the Court didn't draw a map or strike one down on its merits. It told the General Assembly that redistricting, like every other bill it passes, remains subject to the people it governs. Other legislatures currently mid-decade-redistricting on the theory that courts move too slowly to stop them should read the opinion closely. The clock ran to September 2nd this time. It won't always.

By the first week of September 2026, none of the twelve annual appropriations bills has reached the President's desk, and the fiscal year ends in three weeks. This is not a crisis. It is Tuesday. The last time Congress passed all twelve on time, Bill Clinton was finishing his first term, and every September since has produced some version of the same ritual: a continuing resolution, a shutdown threat, a last-minute omnibus nobody read before voting on it, and a great deal of speechmaking about dysfunction from the members casting the vote.

Into that vacuum has walked something with a real name and a real history: impoundment. The Office of Management and Budget, under a director who has been unusually candid about his theory of the office, has leaned on a maneuver called the pocket rescission — sending Congress a request to cancel spending in the final forty-five days of the fiscal year, so that even if Congress does nothing, the money simply expires unspent when the clock runs out. The Government Accountability Office has called versions of this illegal. The administration has proceeded anyway, on the theory that when Congress abdicates the specifics, the executive is entitled to fill in the blanks.

Here is the case for that theory, stated as its proponents actually make it, not as a caricature. Appropriations bills are often written by staff nobody elected, stuffed with riders nobody debated, and passed by members who could not tell you what is in them if their seats depended on it — and sometimes their seats do depend on it, which is worse. If the legislative branch has effectively stopped legislating on spending in any granular, deliberate way, then someone has to exercise judgment about what actually gets spent, and the Constitution puts the person who executes the budget in the executive branch for a reason. An OMB director who cancels a wasteful earmark that Congress never meant to fund with any real scrutiny is not usurping the legislature. He is doing the job the legislature stopped doing.

I understand the appeal of that argument better than I'd like to. My father spent thirty years on a shop floor whose survival depended on decisions made in Washington by people who never had to look him in the eye, and there is something almost satisfying about the idea of one person finally saying no to the machine on the public's behalf. But satisfaction is not a constitutional category, and the argument fails for the reason every argument for concentrated discretion eventually fails: it only works if you trust the discretion, and trust is not a renewable resource in American politics.

The Impoundment Control Act of 1974 was not an accident of Watergate-era overreaction. Richard Nixon impounded funds for clean water projects and highway construction that Congress had appropriated and he had signed into law, on the theory that his judgment about waste superseded the statute. Congress passed the ICA specifically to close that door — to say that if a president wants to cancel appropriated spending, he must ask Congress, in writing, and Congress must say yes within forty-five days, or the money gets spent as written.

That is the whole mechanism. The pocket rescission is not a clever use of that law; it is a way of running out the clock the law was built to prevent.

And the trouble with cheering the maneuver because you like this administration's targets is that the maneuver does not come with a partisan lock on it. The next president who inherits this theory of the office will not necessarily use it to cancel foreign aid line items a lot of voters found dubious anyway. She might use it to zero out a border wall appropriation, or a weapons program, on the same theory: Congress wrote it sloppily, I know better, forty-five days is a technicality. If the power is legitimate when your side holds it, it is legitimate when the other side holds it. That is the entire discipline of constitutional argument, and it is the discipline the current defenders of pocket rescission are asking you to abandon for the length of one news cycle.

None of this excuses Congress. An institution that cannot pass its own budget on time in thirty years has forfeited a great deal of its standing to complain when the executive fills the silence — the same way a family that never balances its checkbook has forfeited some standing to complain about the bank's overdraft fees. But forfeiting standing to complain is different from forfeiting the law, and the ICA remains the law until Congress repeals it, whatever OMB's lawyers currently argue in their memos.

There is an honest fix here, and it does not require anyone to trust anyone's discretion. If OMB believes specific appropriations are wasteful, the ICA already gives it a tool: send the rescission request early enough that Congress has to vote on it, in the open, on the record, with time left on the clock either way. Make members go on record defending the earmark or cutting it. That is not slower government. It is government that has to answer for itself, which is the only kind the founders were interested in building.

Congress abdicated the purse one continuing resolution at a time, and now an OMB director is treating that abdication as an invitation. He may even be right about some of the line items. But the Constitution did not give the executive a line-item veto — the Supreme Court struck down Congress's own attempt to hand him one in Clinton v. City of New York in 1998, on a vote that should embarrass anyone now defending the same power by another name. Put the rescissions on the books, in daylight, on a clock Congress actually has to answer. The alternative is not a leaner government. It is a government where the purse belongs to whoever holds the calendar.

My father kept his Social Security statements in the same drawer as his union card, and once a year he'd lay the new one on the kitchen table next to the old ones and do the arithmetic himself, out loud, distrusting the government's math on principle. He never lived to see the year the trustees now flag as the hinge: 2033, when the combined trust funds are projected to run down to the point that incoming payroll taxes cover roughly seventy-nine cents of every dollar owed. That is not a partisan estimate. It is the number in the Social Security Trustees' own report, the one both parties fund, staff, and then decline to discuss on the same day.

The strongest case against touching any of this is the one seniors' advocates make with more discipline than either party's leadership: that Social Security is a contract already paid for, in payroll deductions going back to a first paycheck, and that a government which can find trillions for banks and airlines and pandemic relief can certainly find the money to keep a promise it wrote into law in 1935 and has never once broken outright. There is real force in this. A retiree who paid into the system for forty-five years did not sign up for a means-tested favor subject to renegotiation whenever Washington gets nervous. She signed up for an entitlement in the oldest sense of the word — a thing she is entitled to, full stop — and the people telling her to expect less are, disproportionately, people who will not feel the cut themselves.

Grant all of that. It still does not answer the question sitting under the drawer where my father kept his statements: what happens on the day the trust fund actually hits zero, under the law as written right now, with no further action by anyone?

The answer is the part neither party wants on a campaign mailer. Current law does not authorize borrowing to cover the shortfall. It does not authorize means-testing on the fly. It authorizes an automatic, across-the-board benefit cut to everyone drawing a check, roughly twenty-one percent, the moment the trust fund can no longer cover the promise. No vote required. No hearing. No senator's name attached. The cut is already law; it is simply dormant, waiting for the calendar to trigger it, and every year Congress declines to act is a year it chooses that outcome by omission rather than owning it by amendment.

Put that on the books where a republic can see it. A benefit cut is a benefit cut whether Congress votes for one in daylight or lets an old formula do it in the dark on a Tuesday in 2033. The Democratic line — that the fix is simply to lift the payroll tax cap and the math resolves itself — is half true and sold as whole; lifting the cap on wages above roughly $176,100 closes a meaningful share of the gap but not all of it, and every serious actuarial model says so. The Republican line — that the fix is raising the retirement age and calling it modernization — is also half true and sold as whole; it saves money precisely by cutting benefits for people who worked with their hands longer than anyone drafting the bill ever will. Neither party will say the plain thing, which is that closing this gap honestly requires some blend of higher revenue, later eligibility, and slower growth in future benefits, in combinations painful enough that nobody wants their name on the vote.

That reluctance is not bureaucratic drift. It is Article I in retreat from its own job. The Constitution gave the power to tax and spend to the branch that has to face voters every two years precisely because that branch was supposed to answer for hard arithmetic in public, not because it was expected to schedule the hard arithmetic for a year when most of today's incumbents will be retired themselves. Every session Congress spends investigating executive overreach while leaving a scheduled twenty-one percent benefit cut untouched is a session spent guarding the wrong flank.

None of this is fixed by a strong executive making unilateral adjustments to disbursement formulas, however tempting that shortcut looks to whichever administration inherits the crisis in real time — and one will inherit it, on a specific date, not as metaphor. The remedy has to run through the branch that can be voted out for the vote itself, which means an actual floor debate on an actual bill, with actual numbers attached to actual names, years before the trigger date rather than in the panic afterward. Seven years is not abundant time, but it is enough, if anyone in either party decided that losing an argument in public was less costly than losing the country's trust by accident.

My father would have wanted the number said plainly rather than managed. Seventy-nine cents on the dollar, arriving on schedule, is not a warning anymore. It is a date on a calendar that both parties have agreed, so far, not to look at.

Ask an American whether the Democratic Party represents the interests of the working class and thirty-seven percent say yes. Ask about the Republican Party and the number drops to thirty. POLITICO published those figures this month. The plurality picked neither.

That plurality is the terminus of a fifty-year drift, not a fresh confusion. A majority of working-class Americans identified as Democrats in 1958. Not since. The share calling itself Democrat has fallen in nearly every decade since Eisenhower, through recessions and expansions, through NAFTA and the 2008 crash, under presidents of both parties. Only in 2020 and 2024, for the first time in the history of the survey, did more working-class voters call themselves Republican than Democrat, and even then the margin was narrow enough that neither party could claim a mandate over the other.

The traditional account of class voting has a long pedigree and a simple mechanism: workers vote their wallets, and whichever party better defends wages and job security wins their loyalty more or less automatically. It is the assumption behind a century of political science, Labour for the working class and the Conservatives for everyone with capital to protect, and it is not a foolish assumption. Wages matter. Job security matters. A worker laid off under a Democratic administration has every reason to blame the party in power, and one who watches a plant reopen under tariff protection has every reason to credit whoever imposed it. If the working class looks volatile, the economic determinist says, that is simply the economy being volatile.

That account explains too little of the record even where it should work best. New Zealand's Labour Party, built explicitly on the industrial working class in a country with one of the cleanest class-party histories in the democratic world, lost a majority of working-class votes to National for the first time in 2011. Historians studying the party's own founding decades, 1911 to 1951, found what they called an "abnormally large population of working-class Tory voters" even then. Class has never been destiny at the ballot box, not in Wellington and not in Dayton. It has been a tendency, weaker than either party's mythology admits.

The more precise account of the American drift is uglier for progressives than "false consciousness" and more specific than "cultural backlash." After 2008, the Democratic Party's explanatory framework shifted rapidly toward structural barriers, systems and institutions beyond any individual's control, and away from the older vocabulary of hard work and earned reward that working-class voters had long used to understand their own lives. Working-class voters did not become more reactionary in response. Democrats changed the terms of the conversation and left a large share of their historic base without the vocabulary to describe its own experience inside it.

Working Families Power's year-long research on the multiracial working class found seven distinct values-based clusters within it, from what they call Next Gen Left to Core MAGA, with several cross-pressured groups sitting uneasily between. Seven clusters inside one "class" is not a rounding error. It is evidence that the thing pollsters label working class no longer votes as a bloc responding to a wage number, if it ever fully did, and that whichever party learns to speak to those clusters on their own terms inherits votes the other side is actively losing.

I grew up going to union meetings with my father in a hall that had a framed photograph of Kennedy on one wall and, by the time I was in high school, an increasingly awkward silence about who the local was actually going to endorse. Nobody in that room had lost a job that year. What they had lost, gradually and without anyone quite naming it, was the sense that either party's headquarters still knew they existed outside of an election year.

None of this means class has stopped mattering, or that the two parties are equally responsible, or that sharper polling language will undo what fifty years of drift produced. It means the working class was never a client base waiting for the correct check to clear. It is a coalition organized around a worldview about earned reward and institutional trust, and the party that keeps mistaking that worldview for a price to be met will keep losing voters it never understood it was representing.