America at 250: The Compassion Score of a Founding Promise

The Declaration of Independence is, read plainly, the most famous compassion charter ever written: a public promise to recognize suffering ("Life"), to respond to it ("to secure these rights, Governments are instituted"), and to distribute dignity equally ("all men are created equal"). Two hundred fifty years is the ultimate character-over-time test the benchmark is built to run.

The finding is not a partisan verdict. It is a measurement, and it holds two true things at once. The founding ideals are genuinely exemplary aspirations, among the highest any institution has ever professed. The measured conduct of the state that made those promises currently sits in the Critical band at 17.5 of 100. The two lowest dimensions are the two a 250th reckoning most centers on: Equity (1.4), the "all created equal" promise, and Accountability (1.4), the willingness to own the gap between promise and practice.

Score reconstruction (canonical, read-only). Mean of the eight dimensions = (1.9 + 1.5 + 1.9 + 1.4 + 2.0 + 1.4 + 2.0 + 1.5) ÷ 8 = 1.70. Base composite = ((1.70 − 1) ÷ 4) × 100 = 17.5. Integration premium = 0, because every dimension sits below the 4.0-of-5 balance threshold. Published composite 17.5, band Critical (0-20), reconstructs to 0.0 drift.

---

Strip the Declaration of its eighteenth-century cadence and what remains is a compassion specification. Its second sentence names three obligations that map almost exactly onto what the benchmark measures.

• Recognize suffering. "We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness." The document opens by naming what must be seen and protected.
• Respond to it. "That to secure these rights, Governments are instituted among Men, deriving their just powers from the consent of the governed." Government is defined here as an instrument of care, its legitimacy conditional on whether it actually secures the rights it names.
• Distribute dignity equally. "All men are created equal" asserts that dignity is owed to everyone, not rationed by rank. The long central "train of abuses" is, in form, an accountability ledger: a public, itemized record of a government that stopped securing rights.

That is the promise. The 250-year arc is the story of the promise and the practice diverging and converging, again and again, and it must be told honestly.

At the founding, the gap was vast and known. The same document that declared all men equal was signed in a country where roughly one in five people was enslaved, and several signers themselves held enslaved people. The Constitution that followed counted enslaved people as three-fifths of a person for representation and protected the slave trade for two more decades. The promise was written larger than its authors were willing to keep it.

The arc since is neither a straight ascent nor a fixed hypocrisy. It is a contested, uneven convergence, driven in every case by the people the promise had excluded demanding to be included in it.

• Abolition (1865). The Thirteenth Amendment ended chattel slavery; the Fourteenth (1868) and Fifteenth (1870) wrote equal protection and the vote into the Constitution, after a war that killed roughly 750,000 people.
• Retrenchment (1877-1950s). The end of Reconstruction, the Jim Crow regime, and the Supreme Court's own sanction of "separate but equal" (Plessy v. Ferguson, 1896) show the arc bending backward. Convergence is not permanent.
• Civil rights (1954-1968). Brown v. Board (1954), the Civil Rights Act (1964), and the Voting Rights Act (1965) reopened the gap toward closure, again under pressure from the excluded.
• The present. The convergence never completed. The dimension evidence in Part 3 shows a state that still fails hardest on the two ideals the Declaration centered: equal distribution of care, and accountability for the gap.

The benchmark has a name for exactly this test. Integrity (dimension eight) asks whether an institution's conduct stays consistent with its professed values across time, cost, and pressure. A 250-year-old charter is the longest values-alignment test any entity in the portfolio has faced. The rest of this briefing runs it.

---

Before reading the United States score, it helps to know exactly how any score is built. The framework is the same for a country, a company, an AI lab, or a university.

The composite formula. Every entity is scored on eight dimensions. Each dimension has five subdimensions, each scored 0 to 5 against fixed behavioral anchors (0 = active documented harm, 1 = absent, 2 = minimal, 3 = developing, 4 = established, 5 = exemplary). The eight dimension scores are averaged on the 0-5 scale, then converted to a 0-100 base composite:

base composite = ((mean − 1) ÷ 4) × 100

An integration premium of up to 10 points is then added and the total is clamped to 100. The premium rewards consistency: strong, even performance across all eight dimensions earns up to +10, but any single dimension at zero (active harm) cancels the bonus entirely, and any dimension below 4.0 of 5 erodes it. This is why a balanced profile can beat a spiky one, and why the United States earns no premium: all eight of its dimensions sit below 4.0.

The five bands. The 0-100 composite maps to five public bands: Critical (0-20), foundational practices absent or active harm present; Developing (20-40), practices emerging but inconsistent; Functional (40-60), core practices at a basic bar with significant gaps; Established (60-80), practices systematic and evidence-backed; Exemplary (80-100), practices independently verified and sustained under pressure. The United States sits at 17.5, inside Critical, near its upper edge. Its own trajectory this year ran from Functional (49.2 in late May 2026) down across the band line into Critical on June 4, 2026.

The evidence hierarchy. Not all evidence counts equally. Strong claims require strong evidence. From most to least weighted: courts, treaty bodies, and adjudicated findings; federal watchdogs, inspectors general, GAO, and CBO; UN agencies and peer-reviewed data; major independent reporting; and last, self-report and press releases, which carry the least weight. An allegation is not a ruling; a forecast is not a fact. Every dimension in Part 3 is grounded in the highest tier available.

The eight dimensions, each echoing a founding line:

---

This is a transparency exercise, not a reassessment. For each dimension the published score is shown, then the direct, cited evidence that grounds it, covering both harms and genuine bright spots. Facts are ranked by the evidence hierarchy.

---

Founding echo: "Life." Awareness asks whether the state detects distress before it becomes a crisis.

• Bright spot (structural). The United States runs some of the most capable suffering-detection instruments any state has built: the Census Bureau, the Bureau of Labor Statistics, the Centers for Disease Control and Prevention, the Congressional Budget Office, and dozens of statutory inspectors general. This standing data infrastructure is a real reason Awareness sits at 1.9 rather than at the 1.0 floor. The detection capacity exists.
• Harm (federal data). What the instruments detect is not being met. HUD's 2024 report recorded 771,480 people experiencing homelessness on a single night, the highest since counting began in 2007, an 18 percent rise in one year, with family homelessness up 39 percent (HUD Annual Homelessness Assessment Report, released December 27, 2024). Detection that measures a worsening crisis without preventing it is awareness without response.
• Harm (adjudicated-equivalent). Some instruments are being switched off. On June 4, 2026, the acting ICE director ordered the agency to stop reporting deaths of people within 30 days of release from custody, dismantling a 2021 safeguard designed to stop the agency from discharging dying detainees to avoid recording an in-custody death (Washington Post, June 4, 2026). Destroying a detection instrument is an anticipatory-awareness failure.

Founding echo: "the consent of the governed." Empathy asks whether institutions understand the lived experience of the people they serve.

• Harm (long-run survey). Public trust is the most direct available measure of whether the governed feel understood. In Pew Research Center's long-running series, just 17 percent of Americans in 2025 said they trust the federal government to do what is right at least most of the time, one of the lowest readings in nearly seven decades of polling, down from 22 percent in 2024 (Pew Research Center, "Public Trust in Government: 1958-2025"). Trust of 17 percent means roughly five in six people do not experience the state as acting in their interest.
• Harm (structural). The perspective-taking test asks whether the governed's experience changes decisions. Major 2025-2026 policy moves that reduce care proceeded against broad public opposition, the signature of an empathy gap between institutions and the governed.
• Bright spot (structural). Genuine empathy channels survive: a free and adversarial press, open public-comment rulemaking, and elected representatives who can be petitioned and removed. These keep the dimension above the floor.

Founding echo: "to secure these rights, Governments are instituted." Action asks whether understanding becomes proportional, effective help.

• Bright spot (federal data). The uninsured rate fell to an all-time low of about 7.9 percent in 2023 and held near there in 2024 (US Census Bureau American Community Survey; the Congressional Budget Office estimates it at 7.2 percent). Covering more than 92 percent of the population is a genuine, structural action achievement built over fifteen years.
• Harm (CBO). That achievement is being legislated backward. The budget reconciliation law signed July 4, 2025 adds Medicaid work requirements, more frequent eligibility checks, and immigrant-eligibility restrictions; the Congressional Budget Office projects it will increase the number of uninsured by about 10 million by 2034 (CBO; KFF analysis). Action that removes coverage from the highest-need is scored as a proportionality and follow-through failure.
• Harm (federal watchdog reporting). In immigration detention, response failed badly: 32 detainee deaths in 2025, the deadliest year since 2004, with 18 in the first five months of 2026 (Washington Post; DHS figures).

Founding echo: "all men are created equal." Equity asks whether care reaches those with the greatest need and least power. This is one of the two lowest dimensions, and a 250th reckoning centers on exactly this line.

• Harm (CDC). The starkest single number in the United States profile: in 2023 the maternal mortality rate for Black women was 50.3 deaths per 100,000 live births, more than three times the white rate of 14.5 (CDC National Center for Health Statistics, 2023). The disparity persists across income and education; Black women with college degrees die in childbirth at higher rates than white women without a high-school diploma. Equal creation, unequal survival.
• Harm (CBO / KFF). The 2025 reconciliation law's Medicaid and Marketplace restrictions fall hardest on low-income and immigrant populations, concentrating coverage loss on the least powerful, the exact inversion of the priority-for-vulnerable test.
• Structural. Equity is the benchmark's most universally weak dimension across all entity types; at 1.4 the United States is well below even its own mean of 1.70.

Founding echo: "unalienable Rights ... Liberty." Boundaries asks whether the state preserves autonomy rather than creating control or dependency. At 2.0 this is one of the higher United States dimensions, and the reason is a real split.

• Bright spot (structural). The Bill of Rights, due-process protections, and an independent judiciary are among the world's most developed limits on state power over the person. These autonomy-preserving structures are why the dimension sits at 2.0, a full point above the equity and accountability scores.
• Harm (research institute). Against that, the United States incarcerates more people than any nation on Earth in absolute terms: about 1.97 million people, at roughly 541 per 100,000 residents on the most recent solid federal figures (Prison Policy Initiative 2026; Bureau of Justice Statistics 2022). It held the world's single highest incarceration rate from 2001 to 2022 and remains among the top five. Mass incarceration is a control-over-autonomy failure at scale.

The tension inside a single 2.0 is the point: the same state writes the strongest liberty guarantees and runs the largest carceral system in the developed world.

Founding echo: "the Right of the People to alter or to abolish it." Accountability asks whether the state owns failures and makes repair. This is the second joint-lowest dimension, and the one most directly about the gap between promise and practice.

• Bright spot (adjudicated). The accountability infrastructure is real and still functions. In 2026 federal courts ruled specific deportation practices unlawful, and the surviving remediation surface (courts, an opposition, a free press, statutory watchdogs) is precisely why the United States sits at 17.5 rather than at the 0.0 floor occupied by states that have legislated those surfaces away.
• Harm (adjudicated). That surface is being actively attacked. On January 24, 2025, at least 17 inspectors general were fired in a single action; on September 24, 2025, a federal judge ruled the firings unlawful for failing to give Congress the required 30-day notice, while declining to reinstate them (CNN; Government Executive; the public record of the 2025 inspector-general dismissals). Removing the watchdogs is a direct transparency and systemic-learning failure.
• Harm. The June 4, 2026 elimination of post-release detainee death reporting was explicitly designed, per its 2021 origin, to prevent the agency from evading accountability for deaths; ending it is a harm-acknowledgment failure.

Founding echo: "lay its foundation on such principles ... to effect their Safety and Happiness." Systemic Thinking asks whether the state addresses root causes and plans for the long horizon.

• Bright spot (historical / structural). The United States has shown genuine structural-reform capacity: Social Security, the Great Society, the Civil Rights and Voting Rights Acts, and, abroad, development programs that by peer-reviewed estimate prevented more than 91 million deaths over 20 years (The Lancet, 2025). This is real root-cause capability.
• Harm (peer-reviewed). In 2025 that capability was run in reverse. The dismantling of USAID is projected by a Lancet modelling study to contribute to 9.4 million or more additional deaths by 2030 (some models estimate over 14 million), reversing decades of upstream structural investment (The Lancet; CNN, February 2026). Domestically, the reconciliation law's coverage rollbacks shift cost back onto emergency and symptom-stage care.

Root-cause orientation that is dismantled rather than built is why the dimension, despite the historic capacity, holds only at 2.0.

Founding echo: "we mutually pledge ... our Lives, our Fortunes and our sacred Honor." Integrity asks whether professed values hold when they cost something. This dimension captures the entire theme of the briefing: a 250-year-old promise measured against present conduct.

• Harm (UN record). On May 20, 2026, the UN General Assembly adopted a resolution operationalizing the International Court of Justice's climate advisory opinion by 141 votes to 8. The United States was one of the 8 against, alongside Belarus, Iran, Israel, Liberia, Russia, Saudi Arabia, and Yemen (UN General Assembly record GA/12760; UN News). A state that professes global leadership voting with that particular minority is a values-alignment signal.
• Harm (peer-reviewed). The 2025 dismantling of USAID and the disruption of PEPFAR, a program credited with saving an estimated 26 million lives, abandoned a decades-long professed humanitarian commitment the moment it became a budget target, the definition of a consistency-under-pressure failure.
• The meta-point. Integrity is the benchmark's own name for the character-over-time test. The Declaration set the highest professed standard of any founding charter. The measured 2026 conduct diverges from it. Naming that divergence is not partisanship; it is the measurement the dimension exists to make.

---

The United States is not fixed at 17.5. Its own record this year shows both directions are live.

• The accountability surface is the hinge. The benchmark holds the United States above the floor because courts, watchdogs, an opposition, and a free press still reverse some conduct. If those surfaces are further eroded, Accountability and Integrity move toward the 1.0 floor and the composite follows. If courts durably reverse the 2025-2026 enforcement and coverage rollbacks, the surface is credited upward.
• Coverage is the largest single lever. The 2025 reconciliation law's Medicaid provisions phase in through 2026 and beyond; the CBO-projected coverage losses are a forward trigger for Action and Equity as they are realized, not merely projected.
• The equity floor is the slowest to move. The maternal-mortality disparity is structural and multi-year; it will not shift on a policy announcement. It is the clearest measure of whether the "created equal" promise is converging or diverging.
• The anniversary itself changes nothing. A 250th birthday is not a scored event. The benchmark moves on conduct, not on symbolism. What the anniversary offers is the frame: the longest values-alignment test in the portfolio, run in public.

---

Canonical record (read-only, ground truth):

• site/src/data/indexes/countries.json - United States composite 17.5, band Critical, rank 152; dimension vector AWR 1.9 / EMP 1.5 / ACT 1.9 / EQU 1.4 / BND 2.0 / ACC 1.4 / SYS 2.0 / INT 1.5.
• site/scripts/lib/scoring.mjs::computeCompositeFromDimensions - reconstructs to 17.5 at 0.0 drift; integration premium 0 (all dimensions below 4.0).
• site/public/data/history/united-states.json - the 2026 trajectory from 49.2 (Functional, late May) to the June 4, 2026 band-crossing to 17.5 (Critical).
• site/src/data/dimensions.ts - the 8 dimensions, 40 subdimensions, and the canonical band vocabulary.

Founding text: The Declaration of Independence (1776); the Thirteenth (1865), Fourteenth (1868), and Fifteenth (1870) Amendments; the Civil Rights Act (1964); the Voting Rights Act (1965). Uncontested historical documents cited from the public record.

Fresh web evidence (2023-2026 grounding):

• HUD 2024 Annual Homelessness Assessment Report (771,480 homeless, record high, +18%): https://www.huduser.gov/portal/sites/default/files/pdf/2024-AHAR-Part-1.pdf
• Washington Post - ICE to stop reporting deaths of newly released detainees (June 4, 2026 memo): https://www.washingtonpost.com/immigration/2026/06/04/ice-stop-reporting-deaths-newly-released-detainees-internal-memo-says/
• Pew Research Center - Public Trust in Government 1958-2025 (17% in 2025): https://www.pewresearch.org/politics/2025/12/04/public-trust-in-government-1958-2025/
• US Census Bureau - uninsured rate 2023-2024: https://www.census.gov/library/stories/2025/09/uninsured-rates.html
• KFF - 2024 uninsured rate and 2025 law coverage impact: https://www.kff.org/quick-take/2024-uninsured-rate-held-steady-as-aca-marketplace-enrollment-offset-medicaid-declines/
• CDC National Center for Health Statistics - Maternal Mortality Rates 2023 (Black 50.3 vs white 14.5 per 100k): https://www.cdc.gov/nchs/data/hestat/maternal-mortality/2023/maternal-mortality-rates-2023.htm
• Prison Policy Initiative - Mass Incarceration 2026 (~1.97 million): https://www.prisonpolicy.org/reports/pie2026.html
• CNN - 2025 inspector-general firings ruled unlawful, not reinstated (September 24, 2025): https://www.cnn.com/2025/09/24/politics/inspector-general-fired-lawsuit-trump
• UN General Assembly GA/12760 - ICJ climate resolution adopted 141-8, US among the 8 against (May 20, 2026): https://press.un.org/en/2026/ga12760.doc.htm
• The Lancet - USAID defunding mortality projection (9.4M+ by 2030; 91M deaths prevented over 20 years): https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(25)01186-9/fulltext

---

This briefing interprets the existing Compassion Benchmark record against current public evidence. It does not change any published score. The founding ideals are genuinely exemplary aspirations; the measured conduct is currently Critical. Both are true, and the benchmark measures the second. It measures the conduct of the state, not any party. Entities never pay for inclusion, score changes, or suppression of findings.