The Capitol Audit Scorecard: We Graded Congress Against Article I. It Got a 51.

The highest grade in the series, and it comes from the chamber built to be the most responsive. Two-year terms and full-body elections (82) and the eligibility bar (90) are untouched since ratification; germaneness rules still keep House amendments tethered to the bill (80); quorum and journal requirements function as ordinary housekeeping (83). The drag is structural rather than textual: the 435-member cap frozen since 1929 while the population tripled (66), a Speaker whose control of the floor determines what the chamber may even consider (48), and the Origination Clause (38) — revenue bills must begin in the House, and the Senate reliably hollows them out by substituting its own text into a House shell. The motion to vacate (62) is the one feature that got measurably more real, having removed a sitting Speaker for the first time in American history. A chamber whose formal design is largely intact and whose agenda-setting has consolidated into one office.

One point behind the House, and the split inside the card is the cleanest illustration of this entire series. Everything about who sits in the chamber scores in the high 80s and 90s — qualifications (93), staggered six-year terms (92), direct election under the Seventeenth Amendment (87). Everything about what the chamber does with its constitutional powers scores far lower: the filibuster (35), which appears nowhere in the Constitution and has been rewritten by whichever party held the votes; the treaty power (28), never repealed but routed around so thoroughly that executive agreements outnumber formal treaties roughly nine to one; the nominations check (55), reshaped by a blue-slip custom both parties invoke selectively. Equal state representation (70) works precisely as written — the grade reflects the strain of a 67-to-1 population disparity the design was never built to absorb, not any deviation from it.

Nine chairs holding the gavels where bills actually live or die, graded on the same four measurements. Five post attendance and legislative output that would rank near the top of any Congress — Grassley highest on the card (80) with 109 bills enacted and 0.5% of votes missed across five decades, Collins (78) with a perfect attendance record since 1997. The other four each carry one specific, sourced flag: Foxx and the for-profit college sector she oversaw (60), Crapo and securities firms (62), Rogers as the single largest House recipient of defense money while chairing the committee writing the defense budget (50), and Jordan (42) — eighteen years, 825 bills sponsored or cosponsored, zero enacted as primary sponsor, while chairing Judiciary. Donor concentration tracking committee jurisdiction is the system working as designed, for better and worse; it is not a crime and this series never printed it as one.

The entry that opened the series, and the one that established the yardstick every individual grading has used since. Republicans averaged 59, Democrats 67.5, and the entire gap traces to attendance and bill counts rather than donors or scandal — the two cleanest records (Thune, 74; Clark, 72) sit in different parties, and so do the two most concentrated donor relationships (Barrasso and energy; Schumer and Wall Street). Not one of the nine carries a criminal charge or conviction. What exists instead is a handful of unadjudicated ethics complaints on both sides of the aisle, reported as exactly that. A middling grade for nine people doing a compromised job with uneven amounts of effort and entanglement — not a scandal sheet, and not a clean bill of health.

The only entry that graded an office's design and its officeholders separately, because the two answers diverge sharply. The design scores 58: no constitutional or statutory basis (30), no codified qualification standard (55), no protection from at-will removal (45) — one person hires, one person can fire, as Trent Lott demonstrated in May 2001. What saves it is the one feature that matters most when something goes wrong: every ruling is advice, and a simple majority can overrule it at any time (100). The nine people who have actually held the post average 80, and the record shows why — the two who got in the most trouble, Dove and Wickham, got there precisely because they ruled correctly and let the chip fall against their own side or their own boss. The remedy has always existed. Congress consistently declines to use it, which is a fact about Congress rather than about the office.

The hinge of the series — the entry where a card stops being mostly-intact and starts being mostly-bypassed. Population equality works close to perfectly (90), and the independent-state-legislature theory was rejected outright in Moore v. Harper (82). Then two entirely separate gaps open at once: a congressional override power under the Elections Clause that has sat unused through the whole modern era (30), and no constitutional limit whatsoever on how often maps may be redrawn (22). Between them sits Rucho (15), in which the Court conceded partisan gerrymanders are "incompatible with democratic principles" and then held itself powerless to police them. Texas and California proved in 2025 what that combination permits, redrawing mid-decade against each other with no census, no election, and no rule broken.

The subpoena power that starts the process rests on solid ground (68), affirmed in McGrain nearly a century ago and narrowed only where it touches a sitting president personally under Mazars. Everything downstream collapses. Criminal contempt (15) runs through the same executive branch Congress is often investigating: ten referrals since 2008, two indictments, and as of this year both of those convictions abandoned by the department that won them. Inherent contempt (12) — arrest by Congress's own Sergeant at Arms, no DOJ or courts required — has been dormant since 1934, and DOJ's own lawyers have now advised Congress not to revive it. Civil suits work on a timeline that outlasts the investigation (38). Executive privilege has become the default first move against nearly any subpoena a White House dislikes (45). Real authority to demand answers; almost none to make anyone give them.

The most lopsided card in the series. The legal wall is the strongest single feature anywhere in the Capitol Audit outside the war clause itself — the Appropriations Clause's exclusivity (90), tested directly in Clinton v. City of New York and never successfully challenged since. Everything depending on Congress using its own power on its own schedule has come apart: regular order at 8, the lowest grade in the entire series, with no on-time bill-by-bill fiscal year since 1997; permanent dependence on continuing resolutions and omnibus packages (20); three funding lapses in twelve months including the longest shutdown in American history (15); a debt ceiling that turns Congress's own enacted obligations into a hostage (22). The one place the executive tried to route around the clause directly, 2025's impoundment fight (30), produced a Supreme Court that ruled both ways within seven months and settled nothing.

The lowest grade in the series, and the clearest single illustration of its thesis. The Declare War Clause itself stands undisturbed at 85 — Madison changed "make" to "declare" deliberately, and no court has ever held a president may start a war alone. Every mechanism built to make that authority operational sits at the bottom: a 1973 statute with a fifty-year unbroken record of never once forcing a withdrawal on its own terms (10); courts that have never issued a merits ruling in a war-powers dispute between the elected branches (10); two authorizations written in 2001 and 2002 for named targets, since stretched across roughly twenty countries (20); and Congress's own reassertion attempts, which have cleared both chambers exactly twice in five decades and died on veto override both times (15). The purse is the one real exception — it ended Vietnam — and even that came with Iran-Contra, where the executive simply financed around it.