I. The problem
The House of Representatives was designed to be the branch of government closest to the people. It is no longer that, and the distance can be measured.
At the first Congress, each member represented about 30,000 constituents. Madison regarded that as near the workable limit, and the founding generation, Federalist and Anti-Federalist alike, agreed that the House must grow with the population. The very first constitutional amendment Congress ever sent to the states, ahead of speech, press, and arms on the list of twelve, was the Congressional Apportionment Amendment, fixing a formula for that growth and settling eventually at one representative per 50,000 people. It fell one state short of ratification in 1792 and, carrying no time limit, remains formally pending today.
For 130 years the House grew after nearly every census, roughly as intended. Then the 1920 census threatened a reapportionment fight nobody wanted, Congress deadlocked for a decade, and the Permanent Apportionment Act of 1929 froze the chamber at 435 members as a way out. No principle was asserted. No founding-era argument was answered. The population has nearly tripled since, and each member now represents roughly 770,000 people, a ratio more than 25 times what Madison defended and higher than any comparable democracy except India.
At that scale, representation is a legal fiction, and the consequences cascade in every direction Americans complain about:
A member cannot know a district of 770,000, so the people with professional access (donors, lobbyists, party leadership, media bookers) occupy the space constituents were designed to fill. Campaigns at that scale cost millions, so fundraising consumes, by credible estimates, 30 to 70 percent of a member's working time and makes donors structurally indispensable. Districts that large are clay for gerrymandering, so most general elections are decided before they are held. Accountability arrives, if ever, at two-year intervals through primaries dominated by small ideological electorates, so members rationally fear their flank and never their middle. Political fame has decoupled entirely from political support, so the loudest members set the agenda while the workhorses labor invisibly. And issues commanding bipartisan supermajorities in the country (congressional stock-trading bans, term limits, drug pricing) die year after year in the building, because small factions hold vetoes the public never granted.
These read as separate pathologies. They are one pathology. Every item on the list is a symptom of a severed feedback loop between the represented and their representatives, and the loop was severed by arithmetic: by district populations that made the founding model of representation physically impossible, and by an institutional structure that has spent a century adapting to that impossibility instead of repairing it.
This essay proposes the repair.
II. The proposal
Four parts, working together.
First, restore the founding ratio. Ratify the Congressional Apportionment Amendment at one representative per 50,000 people. The House grows from 435 members to roughly 6,500.
Second, the portable vote. Any registered voter may reassign the weight of their vote, at any time, through an authenticated government website, to any representative within their own state delegation, and may change it back just as freely. The default, requiring no action ever, is that the voter's weight rests with their district's elected representative, exactly as today. All assignments reset to defaults when each new Congress is seated.
Third, weighted floor votes. On substantive legislation, each member's vote counts in proportion to the voter weight they currently hold. The figure is public and continuously updated.
Fourth, the procedural firewall. Committee assignments, chairmanships, rules, motions, discharge petitions, and quorum remain one member, one vote, untouched by proxy weight. Floor power floats with public support; the institution's machinery does not.
Nothing else changes. The Senate keeps its terms, its equal state representation, and its supermajority thresholds. The presidential veto, judicial review, federalism, and the Bill of Rights are untouched. District general elections proceed exactly as now, on paper, with the winner holding the seat for the full term regardless of any proxy movement. The portable vote operates between elections and on top of them, never in place of them.
A note on lineage, for honesty and for credibility. The expansion half of this proposal is close to consensus among scholars of Congress: the American Academy of Arts and Sciences recommended enlargement, researchers across the political spectrum concur, and the cube-root regularity in comparative legislature size implies an American House far larger than 435. The portable-vote half is newer and less tested. Its nearest formal ancestor is Pivato and Soh's 2020 paper "Weighted Representative Democracy," which proved that a legislature whose members are freely chosen and weighted by their adherents converges, in large electorates, on the outcomes a full referendum would produce. Small delegative-democracy experiments, notably within the German Pirate Party, demonstrated both the concept and a failure mode: delegation concentrated faster than theory predicted. What is original here is the synthesis and the guardrails: the state-delegation cap, the sticky district default, the biennial reset, and the procedural firewall, each aimed at a specific known failure mode.
III. Why it works
The arithmetic of 50,000. A district of 50,000 is a knowable community: a town, a few neighborhoods. A member can meet a meaningful fraction of constituents in person; town halls become real; a campaign is retail, winnable on local standing and volunteer effort rather than millions in donor money. Independent and unconventional candidacies become affordable. And the chamber that results is structurally harder to capture: roughly 12,000 registered lobbyists achieve saturation coverage of 435 members, but covering 6,500 requires an influence industry fifteen times its size, coordinating fifteen times the relationships. Some capture would survive at committee chokepoints, but the wholesale coverage that defines the current system arithmetically cannot.
Coupling power to support. The deepest dysfunction of the current system is that political influence and constituent support have no live connection. A member's power between elections is fixed regardless of anything constituents think or learn. Fame, donor favor, and leadership position all convert to power; demonstrated public trust does not. The portable vote installs the missing wire. A member their constituents trust carries full or amplified weight; a member bleeding trust carries less, visibly, this week. Every unit of the number is a verified voter in the member's own state who took a deliberate, revocable action to confer it. It is the one currency in American politics that cannot be inflated.
Accountability at the speed of information. Under current rules, a scandal on Tuesday meets an ethics referral that concludes in fourteen months and an election up to two years away; the system does not defeat accountability so much as outlast it. Under the proposal, constituents begin reassigning weight within hours, the member's floor power carries the discount within days, and no gatekeeper (leadership, committee, press cycle) can absorb or delay the verdict. Symmetrically, the mechanism forgives: a member who owns a failure and does the repair work can watch weight return, a genuine signal of restored trust rather than mere media fatigue. And because the seat itself remains the district's to give and take at regular elections, weight loss is survivable in a way that preserves room for unpopular judgment. The design intent is deterrence upstream of scandal: penalties that are immediate and probable deter as distant, unlikely ones cannot.
The incentive realignment. Today the activities that secure a career (fundraising, donor cultivation, leadership loyalty, national media) have almost nothing to do with the district, and district contact is accordingly performative. Under the proposal, presence and responsiveness become load-bearing: the member who coasts bleeds weight to the colleague two districts over who shows up, and the differential is public. Seniority flips from pure asset to mixed blessing, since every year in office is another year constituents can drift. Party discipline weakens at the root, because a member ordered to vote against the district now faces an immediate, visible cost for obedience and an immediate, visible reward for refusal. The whip's traditional threats (committee assignments, money) lose force against members who can hold power through constituents alone.
The procedural firewall. Confining proxy weight to substantive floor votes does quiet, essential work. No accumulation of weight can seize a gavel, stack a committee, or manipulate quorum. The unglamorous workhorse with modest weight and deep expertise retains full procedural standing. Two distinct kinds of power emerge, answering to two distinct constituencies: floor power earned from the public, procedural power earned from peers. A healthy legislature needs both, and the current system's error is conflating them.
Gerrymandering, disarmed from three directions. At 50,000 people per district, the surgical shapes become geometrically impossible; districts collapse toward neighborhoods. At 780 districts per large state, the map-drawing optimization becomes intractable to tune and fragile between censuses. And the portable vote attacks the technique's purpose rather than its method: gerrymandering works by wasting votes through packing and cracking, but wasted voters under this proposal simply move their weight to representatives who share their views elsewhere in the state. The map keeps deciding who holds each seat; it largely stops deciding how floor power distributes, which was the point of drawing it crooked. This proposal was not designed to fix gerrymandering. The fix falls out of the design, because gerrymandering is a technique for severing the voter-power link and the proposal exists to keep that link alive.
IV. The Senate, unchanged
This deserves its own section because it reframes every objection that follows.
The proposal touches one mechanism in one chamber. The Senate retains six-year terms, equal state representation, cloture, and every supermajority threshold. This is foundational, not incidental: the Founders built the Senate specifically as the cooling chamber for a hot, responsive House. The famous metaphor pours the House's tea into the Senate's saucer.
Observe the current arrangement through that metaphor: the House, at 770,000-person districts, is too distant from the public to be hot, and the Senate therefore has nothing to cool. Neither chamber plays its designed role; both have converged into a single distant, captured posture. Under the proposal, a genuinely responsive House meets a genuinely deliberate Senate, and any legislation born of a proxy surge still faces the chamber explicitly designed to slow such things, plus the veto, plus the courts. The constitutional checks are not weakened by a responsive House. They are finally activated by one, because for the first time in a century there would be something for them to check in the direction they were built to check it.
V. Objections
"This is direct democracy with extra steps." It is not, and the distinction is precise. Voters never vote on legislation. They choose, continuously, which of their elected representatives speaks for them, which is a purely representative act. Every Madisonian filter (deliberation, committee expertise, the refinement of public views through chosen citizens) remains intact. What the proposal restores is the assumption those filters were built on: that the chosen citizens actually track the public. In the first Congress, that tracking was automatic, enforced by 30,000-person districts and personal acquaintance. Population growth destroyed it; the portable vote rebuilds it with modern tools. This is republican democracy with its feedback loop repaired, closer in operation to 1790 than to the current system, which is the actual historical anomaly.
"It will empower demagogues." Run the numbers with assumptions visible. Most weight never moves: if 10 to 15 percent of voters ever reassign (a generous read of every participation benchmark we have), 85 to 90 percent of floor power rests at its defaults. The state cap closes each pool: no national figure harvests beyond their own delegation. In a deliberately generous scenario, a phenomenon who persuades a million Californians to actively confer their weight reaches perhaps 20 to 40 times a normal member: about 0.6 percent of a 6,500-member floor, held revocably and in public, with zero procedural leverage through the firewall, and the Senate still standing between any House passion and the law. A committee chair today wields more concentrated power, acquired by seniority rather than by a million citizens' active consent. Meanwhile the mechanism does what the current system cannot: it audits the demagogue. Today, claimed armies of followers cannot be checked; under the proposal, the claim is tested continuously, and noise that fails to convert into weight is exposed numerically every day. Demagogues exist now. The proposal does not create them; it prices them.
"If 90 percent never touch it, how does it do anything?" Because political effects scale with margins, not medians, and always have. What disciplines Congress today is primary electorates of 15 to 25 percent, donors amounting to low single digits, activists who are a rounding error: politicians already orient careers around minorities far smaller than 10 percent. The only question is which minority holds the margin, and the proposal replaces donors and primary ideologues with the engaged 10 to 15 percent of general-election voters, 15 to 20 million people, vastly larger and more representative than what it displaces. Deterrence, moreover, does not require firing: most incumbents never lose a primary, yet primary threats discipline all of them, and members would behave differently because weight could move this week, whether or not it does. Low participation even improves the signal, since every movement is a deliberate act by a real constituent, high-grade information no poll matches. Participation is also event-driven: scandals and hot issues produce surges among affected voters at exactly the moments accountability matters. And a concession that strengthens the case: roughly half the proposal's benefits (gerrymandering geometry, lobbying dilution, cheap districts, citizen candidates) flow from expansion alone and require zero proxy usage.
"Real-time pressure will kill political courage." The Burkean objection, and the most serious one. The representative owes judgment, not obedience, and the electoral buffer is what makes the hated-but-right vote survivable; TARP passed because members could be vindicated by events before facing voters. Three answers. The buffer is reduced, not removed: the seat is still decided only at the regular district election, so the courageous vote risks months of diminished weight, not a career, and vindication restores the weight visibly before any ballot. The mechanism also rewards courage as the current system cannot: a member who casts the hard vote and then persuades can watch the persuasion register within weeks, making courage-plus-explanation a viable strategy for the first time. And the honest residual: some marginal courage is still lost, and no design detail eliminates it; the counterweight is that the current buffer mostly protects not statesmanship but two years of consequence-free obedience to donors and whips. The trade is real and worth weighing with both pans loaded.
"It violates one person, one vote." It does not; it nearly perfects it. The doctrine requires each voter to carry equal influence, and under the proposal every voter holds exactly one vote of exactly equal weight; only its holder changes. If anything the current system violates the doctrine in practice, since a gerrymandered voter's equal ballot carries near-zero actual influence, while the portable vote lets it land where it counts. The genuinely open question is different: whether weighted floor voting squares with Article I's 230 years of equal-member practice. The candid answer is that the portable vote likely requires its own amendment, and this essay assumes the harder path. The expansion, by contrast, needs only a statute, since 435 is a 1929 law.
"It will be hacked or coerced." The system is not online voting for office; elections stay on paper, and the portable vote never decides who holds a seat, which removes internet voting's worst cases. Identity verification at this stakes level is what the IRS and Social Security Administration already do daily, and the system can use tools ballot secrecy forbids: full audit logs, anomaly detection, and reversibility, so a fraudulent reassignment, unlike a fraudulent ballot, is detectable and correctable by its victim. Coercion is the harder problem, and the design answer is that assignments are private, receipt-free, and reversible in thirty seconds from any device: a coerced voter complies on Sunday and reverses on Monday, and durable coercion requires durable surveillance, structurally harder against a continuously reversible act than against the one-shot ballots we cast now. The residual risk is real, engineerable, and a cost to weigh, not a disqualifier.
"6,500 members cannot function." Legislatures do not work by chamber-wide conversation; 435 members cannot deliberate together either, and do not. The work happens in committees, the floor ratifies, and the committee model scales comfortably: the UK Commons runs 650 members with fewer chamber seats than members, the European Parliament 720 across 24 languages, the Lok Sabha a nation of 1.4 billion. A 6,500-member House extends the model with hundreds of genuinely specialized subcommittees, against today's members stretched an inch deep across a mile of jurisdiction; there is a respectable case that 435 is the worst of both worlds, too many to deliberate and too few to specialize. Logistics are budget lines. The honest unknown is cultural, how norms and coalition-building evolve at that size, and certainty either way would be salesmanship.
"We cannot afford it." Five to eight billion dollars a year, all in: about a tenth of one percent of federal spending, supervising seven trillion of it. If enlarged capacity improves spending efficiency one percent, the reform pays for itself an order of magnitude over, annually. A cheap legislature that cannot supervise an expensive government is the most expensive arrangement available.
VI. What would follow
Impeachment regains its teeth. The power has become a party headcount because members pay nothing for shielding a president or for impeaching one theatrically. The portable vote prices both behaviors symmetrically and in real time: the shielding member watches constituents withdraw weight during the proceedings; so does the member prosecuting theater the country does not buy. The eventual vote becomes a weighted reading of the country's actual conclusion, formed while everyone watched the meters. The same repricing reactivates contempt, subpoenas, and the purse, with no change to a single word of Articles I or II. The Founders' tools were never broken; the incentive to use them was.
Faction vetoes break. Issues with bipartisan supermajorities die today because coordinated minorities inside each party outweigh uncoordinated majorities outside. Under the proposal, champions of an orphaned popular issue accumulate cross-party weight from voters in their states, comparable champions emerge elsewhere without any coordinator, and the assembled floor weight can force bills out through discharge, procedures the firewall leaves fully intact. The transmission line from durable public consensus to legislative action, currently absent, gets built. Contested issues, where the country genuinely divides, stay contested, and should.
Third parties get a path, on a decades scale. Cheap districts make entry possible for candidates with community standing; the portable vote then amplifies each success, since a lone third-party winner becomes a statewide magnet for every like-minded voter; and the spoiler problem, which polices sincere voting today, dissolves, because a losing sincere vote can still land its weight on a winning sincere representative elsewhere in the state. The two parties are not abolished; they face, for the first time in a century, the competition they organized the system to prevent. Realistically this unfolds over twenty or thirty years, a changed trajectory rather than an overnight multiparty Congress.
Fame stops passing for support. Proxy weight is an audit no politician can inflate: every unit is a verified in-state voter's deliberate, revocable act. The famous firebrand either converts noise into earned, measured, revocable power, or stands exposed daily by the gap between followers and weight. The invisible workhorse finally has a currency in which constituents can pay for unglamorous competence. The prediction, honestly labeled as one, is that a decade of this shifts who runs and who thrives, because the job stops rewarding performance skills and starts rewarding trust.
The job becomes doable by normal people. Every reason your most respected neighbor would never run (millions in fundraising, an unknowable district, a decade of seniority before mattering) is a parameter, and the proposal resets each: retail-scale campaigns, knowable communities, and immediate weight for any freshman who earns real trust. The founding assumption of citizen legislators did not fail because Americans changed; the job's parameters drifted until only a professional class could meet them. An honest caveat: small districts will elect some cranks too, and the bet, stated plainly, is that people chosen by 50,000 neighbors who know them beat, on distribution, people chosen by the fundraising treadmill. That is a wager about human beings, supported by every functioning school board in the country, but a wager.
VII. What could actually go wrong
Arguments for are still arguments for; here is the other ledger, stated as plainly.
The proxy-active minority might be the wrong minority: if reassignment attracts mainly the very online and perpetually furious, the mechanism amplifies intensity instead of diluting it, and this is the design's most important empirical unknown. Delegation might concentrate beyond the scenarios: the Pirate Party experiments pooled power faster than theory predicted, and the guardrails built against that failure are designed-against, not proven-against. Some marginal political courage is genuinely lost to real-time pressure, however favorable the overall trade. The system is untested at scale, and legislators would probe it for exploits with the creativity of people whose careers depend on it; some exploits surface only in operation, which is what the biennial reset and, preferably, a state-level pilot exist to catch. And the transition itself, from 435 to 6,500, would be the largest single change to the House since 1789, with the casualties large transitions have.
Why proceed despite the ledger? Because the status quo's failure modes are not hypothetical. Capture, gerrymandering, unaccountability, and factional veto are observed daily at full scale and compounding annually. The choice is not between a proven system and a risky one; it is between known, worsening failures and designed, bounded, pilotable risks. Sound engineering does not keep loading a structure past its rating because the replacement has not been load-tested. It tests the replacement, and it starts now.
VIII. The path
Stated candidly: this is a twenty-to-forty-year project, and possibly longer.
The sequence that makes sense: the idea circulates in public, in full, with its weaknesses attached, and survives criticism or is corrected by it. A state adopts a portable-vote variant for its own legislature, which states are free to do, and one working example changes the entire conversation. Federal expansion proceeds on its own merits by statute, in whatever increments a redistricting crisis makes thinkable, since it requires no amendment at all. And the federal portable vote follows if and when the pilot record earns it, through an amendment process that is slow by constitutional design.
Long fuses are normal for constitutional ideas: the 27th Amendment waited 202 years, and the Apportionment Amendment is still pending after 237. What matters is that the argument does not age, while the status quo visibly does. Every census widens the ratio, every capture scandal and gerrymander and collapse of trust restates the case, and a proposal aimed at the mechanism rather than at either party's advantage is the rare kind that can wait without spoiling.
IX. The pattern
Step back from the parts and one shape emerges. The current system's accountability is delayed, mediated, and gatekept: it arrives at two-year intervals, filtered through primaries, leadership, donors, and press cycles, each a point where it can be absorbed. The proposal's accountability is immediate, direct, and transparent: it arrives continuously, from constituents personally, in a public number no gatekeeper can intercept. Every benefit argued in this essay, and every risk conceded, is a different face of that single substitution.
The test of a structural reform is whether its benefits cascade from one cause or must be purchased one symptom at a time. Gerrymandering, capture, factional veto, unaccountable fame, the vanished citizen legislator: each is downstream of the same severed feedback loop, and each responds to the same repair, which is what a real cause looks like.
The House was the Founders' answer to the question of how a large republic stays connected to its people. The connection was not lost to malice; it was lost to arithmetic, and then a century of institutional adaptation normalized the loss. What is proposed here is not a new theory of government. It is the old theory, given back its working parts: districts small enough to know, and a people's grip on their own delegated power firm enough to matter between Novembers.
Restore the ratio. Make the vote portable. Let the oldest design in American government finally run at the scale of the country it governs.