The Observatory

The scorecard grades one person at a time. Put the grades together and a picture appears, the same fixed oath applied to 595 sitting officials at once. This is the data, drawn straight from the board. It is offered plainly, for anyone, because a citizen deserves to see the pattern, not just the verdict. Correlation is not causation, and a fixed standard judges individuals, not groups, but the patterns below are real and documented.

How many of our leaders clear the bar

Of 595 sitting, graded officials, 56 clear the support line, 9%. The bar is deliberately high (a credit of 700, roughly a composite of 6.9 against the oath). The largest group is neither the best nor the worst; it is the great competent middle. Almost no one is exemplary, and almost no one is off the cliff. Most simply are not good enough against the standard the seat requires.

Does long tenure breed worse conduct? The board does not show it.

The common worry is that the longer a politician stays, the more entrenched and self-serving they become. The board does not bear that out. Sorted by total years in office, the most senior members score higher, not lower, and show no more office-driven enrichment than newcomers. The likeliest reason is survivorship: those who last decades without disqualifying conduct tend to be the ones who did not have it, and a long record gives more chances to demonstrate the cross-aisle and accountability conduct the standard rewards. The freshman classes score lowest, on thin, untested records. Tenure is not the disease.

By party, and the one event behind the gap

Read this honestly. The party gap is driven overwhelmingly by a single documented event: the December 2020 Texas v. Pennsylvania brief and the related fake-elector effort, which 126 House Republicans signed and which caps each signer. That one act, applied on the fixed bar, pulls the Republican average down. It is a finding about who signed a specific document, not a setting on the scale. Strip that event and the parties land far closer. The standard does not care about the party; it caught a mass act that happened to be almost entirely one of them. It cuts the other way too: the enemy-making flag lands on members of both.

The states, best and worst delegations

Only one state's whole delegation averages a passing grade. The weakest cluster is heavily the same Texas v. PA caps concentrated in a few delegations, the conduct finding again, not geography. See every state · on the map.

The flags

Among the 595 sitting graded officials, 89 carry a confirmed Criterion-8 capping flag (process subversion, the amicus and fake-elector conduct) and 14 a Criterion-10 flag (sustained enemy-making). These are the documented, criterion-class breaches that foreclose support no matter the number. The board also holds 170 former and historical figures, kept as the record of who has held power and the seed of the analysis of how the same few keep returning.

Concentration of power: the same few

Ask why the same names keep returning, and the data answers in two ways. First, sheer duration: across the board, 127 people have held office for 20 years or more, 56 for over 30. The longest-serving sitting members have each held federal office for roughly half a century. The seats do not turn over; the people in them age in place.

Second, family. Most shared surnames are coincidence, but some seats pass within a bloodline. These are the documented political families on the board, where the seat moved from one relative to the next:

Put it together with the two-party gate: when two private organizations control who can run, when incumbency runs for decades, and when seats pass within families, the result is the narrow, recycled political class the data shows. The standard does not judge anyone for their name or their tenure; it judges conduct. But the pattern of who keeps the seats is its own finding, and the people deserve to see it.

The governing few, across all of history

Widen the lens to the whole American experiment and the picture sharpens. Since Congress first met on March 4, 1789, a total of 12,591 individuals have ever served in it, 11,262 in the House and about 2,018 in the Senate. The House itself has been capped at 435 seats since 1929, no matter how the country grew. Roughly twelve thousand people, across two and a half centuries, have written the federal laws for everyone else. House Historian (Jan 2026)

And the seats have always run in part through blood. The definitive study (Dal Bó, Dal Bó & Snyder, 2009) found that about 1 in 11 of all legislators in U.S. history had a relative who served before them, a share that ran above 15% in the founding decades, settled near 11% through 1858, and eased to about 7% after 1966, higher in the Senate than the House throughout. One family alone, the Breckinridges, placed seventeen members in Congress. The recycled political class is not new; it is the oldest pattern in the institution. Dal Bó et al., "Political Dynasties"

Meanwhile the people in the seats keep getting older and staying longer. The 119th Congress, at an average age of 58.9, is the third-oldest in history; the Senate's median age is 64.7. Incoming members now arrive with roughly 9 to 11 years of prior service, against the 2 to 3 years that was normal in the 1800s. Turnover has not kept pace with the country. Pew (119th Congress)

Now set that against the governed. Over the same span, an estimated well over half a billion Americans (~550 to 600 million) have lived in the United States, of whom roughly 210 to 260 million have already lived and died, with about 342 million alive today. Hundreds of millions of people, born, working, paying, serving, dying, all of them governed by a legislature that at any moment seats 535, and that in all of history has been filled by about twelve thousand names. That is on the order of one federal legislator for every forty to fifty thousand Americans who have ever lived, and a meaningful share of those few were related to each other. The standard judges what each one did with the seat. The arithmetic of who ever got to hold it is its own quiet verdict. PRB method · Census · cumulative figure is a demographic estimate (±15%)

The representation gap: a different class

Who governs is not, on the numbers, much like who is governed. This is context, not a grade. The standard does not penalize wealth: where it came from honestly is no one's business, and only office-driven enrichment is ever scored as a breach. But the distance between the governing class and the governed is itself a fact worth seeing, because a body that no longer resembles the country it speaks for will, over time, stop hearing it.

Begin with the one comparison that carries no estimate. A rank-and-file member of Congress is paid $174,000 a year, frozen at that level since 2009. The median American household earned $80,610 in 2023. The floor of congressional pay is more than double the middle of American life, before a single investment. CRS, Salaries of Members of Congress · Census, Income 2023

Net worth widens the gap, though those figures must be read with care (see the caution below). The median member of Congress was worth about $1 million in 2018 and roughly $1.28 million by 2021; the median American household was worth $192,900 in 2022. Depending on the years paired, the governing class holds something like five to twelve times the wealth of the governed. And more than half of Congress are millionaires, against fewer than one in ten American adults. The chamber that writes the tax code, the labor law, and the health system is, by majority, drawn from the top of the distribution it legislates for. OpenSecrets · Federal Reserve, SCF 2022

At the top the distance turns abstract. The wealthiest member is worth several hundred million dollars; dozens are worth tens of millions. The richest seat is worth more than a thousand median American households put together. And the gap has widened: Congress first became majority-millionaire around 2012, and through the financial crisis, from 2007 to 2010, typical American wealth fell by roughly 40 percent while the median member of Congress grew richer. The representatives and the represented have been moving in opposite directions. Ballotpedia, Personal Gain Index

A necessary caution on the wealth figures. Congressional net worth is estimated from financial-disclosure ranges and excludes both salary and a primary residence, so it is not measured with the rigor of the Census and Federal Reserve figures for ordinary Americans, and the multiple swings with the years and method chosen. Read the wealth numbers as direction and order of magnitude, not precision. The salary-against-income line is the only one here free of any estimate, and it is the one to lean on.

Two kinds of service: the cost, and who carries it

A final perspective, and the heaviest. War is policy, and policy is never scored on this site, in either direction. But the governing few do one thing no other office does: they vote to send other people to die, and they rarely die themselves. Both parties have cast these votes, most of them near-unanimously. Set the one kind of service, holding a seat, against the other, serving in the war that seat ordered, and the disproportion is the whole point.

About 12,591 people have ever served in the Congress that declares and funds America's wars. In those wars, roughly 1.3 million Americans in uniform have been killed and about 1.5 million more wounded. That is more than a hundred killed, and more than a hundred wounded, for every single person who has ever held a seat in the body that ordered it. A few hundred vote yes; hundreds of thousands carry it. CRS RL32492, American War Casualties · VA, America's Wars

"Voted yes" = total House plus Senate members who voted for the declaration or authorization. Casualties are U.S. military, per CRS / VA. The Civil War, which had no single authorizing vote, killed at least 364,511 in Union service alone, the deadliest war in American history.

And the people casting those votes increasingly have no idea, from the inside, what they are ordering. In the early 1970s roughly three in four members of Congress were military veterans. Today it is fewer than one in five. The 2001 and 2002 votes that launched the longest wars in American history were cast by a Congress already down near that modern low. The hand on the lever is, more and more, a hand that never served. Pew, veterans in Congress

It is worth knowing how few even paused. Jeannette Rankin cast the lone vote against entering World War II. Wayne Morse and Ernest Gruening were the only two against the Gulf of Tonkin Resolution that opened Vietnam. Barbara Lee cast the single vote against the 2001 authorization that is still in use today. The standard this site holds is conduct, not the war vote, which it does not grade. But a structure that lets a few hundred, mostly untouched, spend the lives of millions is the closed system at its most consequential, and that cost has only ever run one direction.

These figures count deaths in uniform during the wars themselves. The fuller toll runs higher: Brown University's Costs of War project counts more than 30,000 post-9/11 veteran and service-member suicides, over four times the number killed in post-9/11 combat. The bill for a vote does not close when the war does. Costs of War, Brown University

Opportunity versus ability: the math of a closed system

The deepest question is not who governs, but whether anyone else realistically could. Set the ability to serve against the opportunity to serve, and the system's one-sidedness becomes arithmetic. Roughly about 200 million Americans are eligible to hold a House seat (citizens over 25). In any given election, only about 40 of the 435 seats are realistically winnable by a newcomer: open seats with no incumbent (8 to 12 percent of races) plus the handful that actually flip. The rest are held by incumbents who win re-election about 95 percent of the time, in districts that are roughly 90 percent "safe."

ACCESS RATIO = eligible citizens / winnable seats ≈ 200,000,000 / 40 ≈ 5 million Americans for every realistically open seat.

That is the gap between the right to serve and the chance to. And it has tightened. Define a transparent Closure Index, the average of the three locks that keep a seat from changing hands, each running 0 (wide open) to 1 (shut):

C = ( R + S + T ) / 3
R = incumbent re-election rate · S = safe-seat share · T = 1 − turnover rate

Inputs are illustrative estimates from the cited literature: re-election and turnover from Brookings Vital Statistics and Ballotpedia; safe-seat share from Mayhew (1974) and FairVote. The index is a transparent composite, not a natural constant.

In one working lifetime the seats went from largely shut to almost sealed. The collapse of competitive districts is the engine: the share of genuinely contested House seats fell from about 30 percent in the early 1970s (Mayhew's "vanishing marginals") to roughly 8 percent today. And the doors themselves are too few for the crowd: each seat now stands for 761,000 people, against 34,000 at the founding, a 22-fold dilution, with the House frozen at 435 since 1929 even as the population tripled. By the cube-root law that fits most world legislatures, a country this size would seat roughly 692; the House is at 63 percent of that. Mayhew (1974) · FairVote · apportionment · cube-root law

Left: U.S. apportionment data (each seat now carries 22 times the founding load). Right: Mayhew's marginals (1962, 1970) and FairVote (2024); fewer doors, and the few that remain rarely open.

None of this is invented. The economists Dal Bó, Dal Bó and Snyder showed that "political power is self-perpetuating," that the longer one holds a seat the likelier a relative follows. Gehl and Porter model the two parties as an industry duopoly with the highest barriers to entry of any market. The consequences are documented too: safe seats produce more extreme legislators (Kustov), lower competition raises corruption (NBER), and the landmark Gilens and Page study of 1,779 policy decisions found that policy tracks economic elites and organized interests while "average citizens have little or no independent influence." Meanwhile Congress earns about 10 to 17 percent approval (it has fluctuated there across 2025 and 2026) against about 95 percent re-election, the paradox of a public that loathes the institution and is given no real way to change it. Dal Bó (2009) · Gilens & Page (2014)

The Closure Index is a constructed measure, offered transparently with every input sourced; treat it as a clear way to see documented facts together, not a natural constant. But the facts it gathers are not in dispute. The standard on this site judges what each officeholder did with the seat. This is the prior question: how few are ever allowed near one, and why the same names keep filling them.

Population: 595 sitting officials with a verified grade (of 640 graded board-wide). Conduct only; policy is never scored. Correlation is not causation; confounders are real and disclosed; the standard judges individuals, not groups. Drawn live from the board. See the methodology.