For 225 years, the United States traced a path through Liberty×Capability space that was, by the standards of comparative politics, almost monotonically northeast. The country began with a Liberty score of roughly 42 in 1800 — a slaveholding republic with a restricted franchise — and climbed, through civil war, reconstruction, suffrage expansion, civil rights legislation, and institutional deepening, to a peak of L=94 around 2015. It was not a straight line. There were retreats during the Red Scares, setbacks during the Gilded Age, and long decades when the promise of liberty was extended to some Americans, and violently denied to others. But the trajectory was unmistakable. Decade by decade, election by election, the arc bent towards inclusion.
Then something changed.
The decline that began in the late 2010s and accelerated sharply after 2024 represents the fastest horizontal collapse in the Political Topology dataset. No other established democracy — defined as a country that maintained L≥80 for 25 or more consecutive years — has declined this rapidly without a military coup or foreign invasion. The comparison set is empty. When we look for historical analogues, we find them not amongst peer democracies but amongst hybrid regimes, and countries undergoing revolutionary change.
In September 2025, the Varieties of Democracy Institute (V-Dem) at the University of Gothenburg reclassified the United States as an "electoral autocracy" — the first time a G7 nation has received that designation whilst its economy remained in expansion. The Century Foundation's Democracy Metre independently scored the US at 57, placing it in the "crisis zone." Freedom House, which grades on a more forgiving curve, and gives significant weight to electoral processes, maintained the US at 84 but flagged the fastest rate of decline it has ever recorded for a country in its "Free" category.
The disagreement between indices is not a bug. It is the most important feature of the data. Different indices measure different things, weight different components, and update on different timescales. Understanding what each index captures — and what it misses — is essential to any honest assessment of where America stands.
The question is not whether the United States is declining on democratic metrics. Every serious measurement agrees on that. The question is how far the decline has gone, how fast it is moving, and whether the structural advantages of a 248-year-old democracy at $85,000 GDP per capita provide a floor that the cross-national model does not capture.
There is a 36-point spread between the most alarming estimate (L=48 on the Political Topology Index) and the most conservative (L=84 on Freedom House). That spread is not noise — it reflects fundamentally different methodological choices about what "democracy" means and how quickly institutional damage should be reflected in a country's score.
The PTI, which updates on a 2-year rolling window and heavily weights regulatory independence and civil society indicators, captures the most recent institutional changes — the Schedule F reclassifications, the inspector general firings, the regulatory agency restructuring — and prices them immediately. Freedom House, which conducts annual expert surveys with a longer institutional memory, and gives substantial weight to the existence of free elections, updates more slowly, and implicitly assumes that the 2024 election's procedural integrity provides a floor under the score.
Neither approach is wrong. They are measuring different things on different timescales. The PTI asks: "What does the institutional landscape look like right now?" Freedom House asks: "How does this country's overall democratic quality compare to the global distribution?" V-Dem falls somewhere in between, with its mid-range scores of 65–70 reflecting a blend of rapidly updated expert assessments and more stable structural indicators.
The responsible analytical move is not to pick a number. It is to show what follows from each number, and to identify which conclusions are robust across the entire credible range.
This is the central analytical contribution of the stress test. Rather than arguing for a single Liberty score, we present the full spectrum of credible estimates, and show how each downstream conclusion changes across the range. The reader can anchor to whichever index they find most credible and trace the implications.
| Metric | L=48 (PTI) | L=57 (TCF) | L=65 (V-Dem Low) | L=70 (V-Dem Mid) | L=75 (Cross-Index) | L=84 (FH) |
|---|---|---|---|---|---|---|
| Stage | S6: Soft Dictatorship | S5: Electoral Autocracy | S4: Competitive Authoritarianism | S3: Democratic Erosion | S3: Democratic Erosion | S2: Early Warning |
| Velocity | −18.0/yr (2yr) | −3.7/yr (10yr) | −2.9/yr (10yr) | −2.4/yr (10yr) | −1.9/yr (10yr) | −1.0/yr (10yr) |
| Event Horizon | BELOW | BELOW | Above | Above | Above | Above |
| Hist. Reversal to L≥70 | 54% | 71% | 86% | 82% | 91% | 99% |
| Predicted Yield | 16.2% | 13.1% | 10.3% | 8.5% | 6.8% | 3.6% |
| Narrative | Critical instability | Crisis zone | Serious erosion | Declining democracy | Declining democracy | Stressed but intact |
Stage classification shifts dramatically across the range. At the PTI's L=48, the US is classified as a "soft dictatorship" — a regime that maintains the trappings of democracy while having effectively neutralized independent checks on executive power. At Freedom House's L=84, it is a "stressed but intact" democracy experiencing its first serious institutional test. The difference between these two readings is not academic. It implies entirely different policy responses, risk assessments, and historical analogues.
Velocity is perhaps the most sensitive metric to methodological choice. The PTI's −18.0/yr reflects its 2-year rolling window capturing the post-2024 acceleration. The 10-year averages used by other indices smooth this signal substantially, producing velocities of −1.0 to −3.7/yr. Both measurements are "correct" — they simply describe different time horizons. A doctor measuring a patient's heart rate during a sprint will get a very different number than one measuring a resting average, but neither is lying.
Event Horizon is binary: either the US has crossed the threshold below which internal self-correction becomes historically improbable (requiring external pressure or elite rupture), or it has not. The PTI and TCF place the US below this line. V-Dem, the cross-index composite, and Freedom House place it above. This is the single most consequential disagreement in the table.
Historical reversal rates provide grounds for both alarm and reassurance. Even at L=48, 54% of countries in the historical dataset eventually returned to L≥70. At L=75, the figure is 91%. But "eventually" can mean 20 or more years, and the post-2006 structural break means that contemporary reversal rates are substantially lower than the full historical average.
Cross-national models treat institutions as a single composite. But the American system is a federation of institutions with wildly varying levels of resilience. Some are holding. Others have already been compromised. Understanding which is which matters more than any aggregate score.
The U.S. military remains the most resilient institution in the democratic architecture. The officer corps' institutional identity is built around oath-to-Constitution rather than oath-to-leader, a tradition reinforced by the post-Vietnam professionalisation reforms. Joint Chiefs Chairman turnover has followed normal patterns. The military's refusal to participate in the January 6 crisis and its subsequent institutional distancing from partisan politics are historically significant. The 20-point deduction reflects increasing politicisation of defence appointments and rhetorical pressure on military leadership, but the core institutional culture remains intact.
The Fed's operational independence is legally protected and institutionally deep. Interest rate decisions continue to follow internal models rather than political directives. However, the rhetorical assault on Fed independence — public demands for rate cuts, threats to the Chair's tenure, and proposals to restructure the Board — has created a chilling effect. Market participants increasingly price in a "Fed capitulation" premium. The 35-point deduction reflects not what has happened but the probability-weighted risk of what could happen. A Fed that loses credibility is a financial system without an anchor.
The federal judiciary presents a paradox. Its structural independence — life tenure, salary protections, Article III status — remains constitutionally intact. No judge has been removed for political reasons. But the ideological composition has shifted dramatically through legitimate appointment processes, and judicial rulings have increasingly aligned with executive preferences on separation-of-powers questions. The Supreme Court's decisions on administrative law, executive immunity, and regulatory authority have systematically expanded executive power. The courts still function. The question is whether they function as a check on power or as a validator of it.
Congress is the weakest link in the institutional chain. The House's oversight function has been largely subordinated to partisan objectives. Subpoena power has been wweaponised against political opponents rather than directed at executive accountability. The Senate's confirmation process has become a rubber stamp for executive nominees. Cross-partisan coalitions — the historical mechanism for congressional independence — have effectively ceased to function. Appropriations, the ultimate congressional power, have been ceded through continuing resolutions, and executive impoundment. The institution exists. Its independence does not.
American federalism is a democratic redundancy that has no equivalent in most countries experiencing erosion. When Hungary's central government captured the judiciary, there was no state-level judicial system to provide an alternative. When Turkey's parliament was subordinated, there were no state legislatures maintaining independence. The US has 50 state governments, many of which are actively resisting federal overreach through litigation, regulatory divergence, and institutional non-compliance. California, New York, and other large states have effectively become parallel governance systems. The weakness is that federal preemption can override state resistance, and that many state governments are experiencing their own erosion dynamics.
American civil society — the constellation of NGOs, foundations, universities, professional associations, churches, unions, and civic oorganisations — remains amongst the densest and most resourced in the world. But the space is shrinking. Tax-exempt status has been wweaponised against oorganisations perceived as politically hostile. Universities face funding threats tied to political compliance. The ACLU, once a consensus institution, now operates in an environment where legal advocacy itself is treated as partisan. The depth of American civic life is a genuine structural advantage. Whether it can withstand sustained institutional pressure is the open question.
A stress test that does not take counter-arguments seriously is not a stress test. It is advocacy. What follows is a genuine attempt to evaluate the strongest objections to the declining-democracy thesis, rated by analytical strength. The reader will note that three of the seven counter-arguments are rated "Strong" — and that this assessment is intentional.
The argument: The United States has 248 years of continuous democratic governance, the world's oldest written constitution, and institutional depth that no other country in the backsliding dataset possesses. The military scores 80. The Fed scores 65. These are real sstabilisers, not theoretical ones.
Assessment: This is a serious argument. The institutional scorecard confirms that two critical institutions — the military and the central bank — remain substantially intact. The cross-national model, built largely from countries with 30–70 years of democratic experience, may systematically underestimate the resilience of a 248-year-old system. However, institutional resilience is not the same as institutional immunity. Hungary's judiciary was also considered resilient. Poland's constitutional tribunal had strong traditions. The question is not whether American institutions are stronger than average, but whether they are strong enough to withstand the specific pressures they currently face.
The argument: The most recent presidential election was conducted without significant fraud, the result was accepted by all major parties, and power was transferred peacefully. How can a country be an "electoral autocracy" when its elections work?
Assessment: The procedural integrity of the 2024 election is not in dispute. But the thesis is about trajectory, not current state. V-Dem's "electoral autocracy" classification rests not on election day fraud but on the broader institutional environment: media capture, regulatory independence, judicial alignment, and civil liberties constraints that make the playing field uneven even when the voting itself is clean. Hungary holds elections too. So does Singapore. The existence of a fair election is necessary but not sufficient for democracy. More importantly, the question is whether the next election will be conducted under the same conditions.
The argument: Democracy indices are produced by Western institutions with liberal-democratic priors, and they systematically ppenalise countries that deviate from the Western liberal model.
Assessment: If Freedom House has a pro-Western bias — and there is a reasonable case that it does — then it overstates the US score, not understates it. The bias argument, if accepted, makes the decline more alarming, not less. A biased-in-favour-of-America index still recording America's fastest-ever decline is a stronger signal, not a weaker one. This counter-argument, logically followed to its conclusion, supports the thesis it is meant to undermine.
The argument: If democratic erosion posed real economic risks, financial markets would reflect it. US equities are near all-time highs. The dollar remains the world's reserve currency. Treasury yields are not pricing in sovereign risk. The market disagrees with the thesis.
Assessment: The historical lag between institutional erosion and market repricing is 3–12 years. Turkish equities were near all-time highs in 2013, three years into Erdogan's consolidation. Hungarian markets performed well through 2015. Brazilian markets rallied during Bolsonaro's first year. Markets price cash flows and discount rates. They are not designed to price institutional quality. Moreover, the "Great Decoupling" — the divergence between democratic quality and economic performance — is one of the most robust findings in the dataset. Markets and democracies measure different things. The absence of market panic is not evidence of institutional health.
The argument: Much of what is being measured as "democratic erosion" is actually policy reversals that the measurer disagrees with. Deregulation is not autocratisation. Judicial appointments are not court-packing. Immigration enforcement is not civil rights erosion. The democracy indices are smuggling in policy preferences under the guise of institutional measurement.
Assessment: This is the strongest methodological objection and it deserves serious weight. There is a real danger of conflating "policies I oppose" with "democratic backsliding." The line between a legitimate policy reversal and structural institutional damage is genuinely difficult to draw, and reasonable people can disagree about where it falls. However, the distinction is not impossible to draw. Firing inspectors general across multiple agencies is not a policy choice — it is the removal of an oversight mechanism. Reclassifying civil servants under Schedule F is not deregulation — it is the politicisation of the administrative state. The test is whether the action is reversible by the next administration or whether it permanently degrades institutional capacity. Not every policy change is erosion. But not every act of erosion presents itself as one.
The argument: Amongst countries that have maintained L≥80 for 25 or more years, 98% of decline episodes eventually reversed. The US has maintained democratic governance for 248 years. The base rate for permanent collapse in mature democracies is effectively zero.
Assessment: The historical record strongly supports this counter-argument. Long-standing democracies have deep institutional roots, democratic political cultures, and self-correcting mechanisms that younger democracies lack. The 98% reversal rate for established democracies is a genuine finding, not an artifact. The question is whether the current episode is within the distribution of normal democratic stress-testing or represents a tail event. The honest answer is: we do not know yet. The base rate says recovery is overwhelmingly likely. But the velocity of decline is unprecedented in the dataset, which means we are, by definition, in territory where base rates may not apply.
The argument: No democracy above $15,000 GDP per capita has ever collapsed into sustained autocracy. The United States has a GDP per capita of approximately $85,000. The economic foundations of American democracy are an order of magnitude above the danger zone. Wealth creates a middle class with stakes in institutional stability, and that middle class is a democratic anchor.
Assessment: The "Lipset threshold" is one of the most robust findings in comparative politics. The correlation between wealth and democratic stability is strong, persistent, and not obviously confounded. At $85,000 GDP per capita, the US is so far above the historical danger zone that the comparison barely applies. The weakness of this argument is that it may be backwards — perhaps high-GDP democracies have never collapsed because none has ever been seriously tested, not because wealth provides actual protection. We may be about to find out whether the Lipset threshold is a causal relationship or a statistical coincidence. But the burden of proof is on those who claim that this time is different.
Taking the counter-arguments seriously does not eliminate the concern. It recalibrates it. The original thesis — that the US is in a "critical instability zone" at L=48 with a 12% reversal probability — almost certainly overstates the danger. But the most generous reading — that the US is a "stressed but intact" democracy at L=84 with a 99% reversal rate — probably understates it.
The intellectually honest position is somewhere in the middle, probably in the L=65–75 range: a country experiencing serious democratic erosion, with genuine structural advantages that the cross-national model underweights, but facing unprecedented velocity of decline that those advantages have not yet been tested against.
The search for historical analogues to the American situation is both irresistible and treacherous. Every comparison illuminates something; every comparison distorts something else. The most commonly invoked parallels range from the instructive to the misleading.
The French Fourth Republic collapsed in 1958 under the weight of colonial crisis, governmental instability, and a military that was no longer willing to support civilian authority. Charles de Gaulle was summoned from retirement to create the Fifth Republic — a new constitutional order with a dramatically strengthened presidency. The transition was democratic in form (approved by referendum) but authoritarian in substance (de Gaulle governed by decree for months, rewrote the constitution to his specifications, and subordinated the legislature).
The analogy is instructive because France in 1958 was a wealthy, established democracy with deep institutional traditions that nonetheless underwent a fundamental constitutional transformation. The Fifth Republic ultimately recovered its democratic character, but the recovery took decades, and required de Gaulle's own resignation in 1969. The lesson: democratic crisis in a wealthy, institutionally deep country can lead not to permanent autocracy but to a constitutional reboot — a painful, disorienting transformation that preserves democratic DNA even as it rewrites the constitutional code.
If the French analogy holds, the United States may be heading not towards dictatorship but towards a Sixth Republic: a transformed constitutional order that concentrates executive power, weakens legislative independence, and redefines the relationship between federal, and state authority. Such an outcome would be profoundly disruptive without being permanently autocratic.
The Weimar analogy is emotionally powerful and analytically weak. Weimar Germany was a young democracy (14 years old) with no democratic tradition, facing economic catastrophe (hyperinflation, then depression), with a military that was actively hostile to the democratic order, in a society that had never developed democratic political culture. The United States in 2026 is a 248-year-old democracy with the world's largest economy, a military that scores 80 on institutional resilience, and a deeply embedded (if currently strained) democratic culture.
Every factor that enabled Weimar's collapse is absent in the American case. The comparison is not just inaccurate; it is counterproductive, because it encourages a binary frame (democracy vs. fascism) that obscures the more likely outcome: gradual institutional degradation within a formally democratic framework. The danger is not 1933. It is competitive authoritarianism on the Hungarian model — a system that holds elections, maintains a market economy, and preserves civil liberties for the compliant while systematically disadvantaging opposition.
When we look at the full historical record of countries that declined to the L=45–55 range, the picture is cautiously encouraging but not comforting:
| Starting Band | % That Recovered to L≥70 | Median Recovery Time | % Still Below L=70 After 20 Years |
|---|---|---|---|
| L = 75–85 | 91% | 6 years | 4% |
| L = 65–75 | 82% | 11 years | 9% |
| L = 55–65 | 71% | 14 years | 16% |
| L = 45–55 | 54% | 18 years | 28% |
| L = 35–45 | 31% | 22+ years | 47% |
Seventy-one percent of countries at L=45–55 eventually recovered. But "eventually" can mean 20 or more years — a generation of degraded institutional quality, constrained civil liberties, and reduced democratic accountability. And these are global averages that include both countries with and without the structural advantages the US possesses. The US-specific factors (wealth, democratic tenure, institutional depth) probably push the recovery probability towards the higher end of the range for any given Liberty score. But they do not guarantee rapid recovery, and they do not prevent the damage that occurs during the erosion period.
The Political Topology thesis was subjected to an independent methodological audit that examined 12 core claims. The results were mixed in the most productive possible way: the thesis was neither vindicated nor demolished, but substantially recalibrated.
| Claim | Status | Detail |
|---|---|---|
| Dataset construction | Confirmed | V-Dem, Freedom House, and supplementary indices properly compiled; 91 countries, 225 years, 1,656 country-year observations verified |
| Yield–Liberty relationship | Confirmed | Negative correlation between Liberty score and sovereign yield spread is statistically significant and economically meaningful across multiple specifications |
| Great Decoupling | Confirmed | Post-2016 divergence between US democratic quality and economic performance is real and unprecedented in the dataset for an established democracy |
| Velocity cataloging | Confirmed | US decline velocity, regardless of which index is used, is the fastest amongst countries that maintained L≥80 for 25+ years |
| Claim | Status | Detail |
|---|---|---|
| Bistable dynamics | Refuted | The claim that democracies exhibit bistable equilibria (stable democracy or stable autocracy with an unstable middle) is not supported by the data. Transitions are gradual, not catastrophic |
| Shock sigma (σ=3.2) | Refuted | The US decline is not a 3.2-sigma event. Proper baseline calculation yields 1.8–2.4 sigma depending on index — unusual but not unprecedented |
| Markov property | Refuted | Democratic trajectories are strongly path-dependent; future states depend on history (democratic tenure, institutional depth), not just current Liberty score |
| L=48 as established fact | Refuted | L=48 is one of six credible estimates spanning a 36-point range. Presenting it as "the" US Liberty score was misleading |
| Original tyranny probability | Refuted | The initially published probability of "sustained autocratic outcome" was based on the refuted Markov model and the refuted L=48 point estimate. Recalibrated probabilities are substantially lower |
| Claim | Status | Detail |
|---|---|---|
| Event horizon concept | Partially Valid | A nonlinear drop in reversal probability does exist in the data, but the threshold is a range (L=50–60), not a sharp line, and it is modulated by democratic tenure, and wealth |
| Velocity measurement | Partially Valid | Velocity is real and significant, but highly window-dependent. The 2-year window that produces −18.0/yr is defensible but misleading without context; 10-year windows give −1.0 to −3.7/yr |
| Treasury mispricing | Partially Valid | A yield-Liberty gap exists and is historically anomalous. But labelling it "mispricing" assumes market inefficiency without sufficient evidence. The 3–12 year repricing lag means the market may be rationally discounting future recovery |
The audit forces a fundamental shift in framing. The original thesis presented the US situation as a crisis with a deterministic trajectory. The recalibrated version presents it as a serious and unprecedented stress test with a probabilistic range of outcomes. The first framing is more dramatic. The second is more honest. And honest analysis, in a democracy under stress, is itself a form of institutional resilience.
The five refuted claims are not embarrassments to be mminimised. They are the mechanism by which the analysis improves. A thesis that survives an audit unchanged was either trivially true or inadequately tested. The Political Topology thesis was neither. It made bold claims, several of which were wrong, and the correction makes the surviving claims more credible, not less.
The value of a stress test is not in its point estimate but in the forward-looking indicators it identifies. What follows are the four domains that will determine, over the next decade, whether the American exception proves to be a temporary deviation, or a permanent departure from the democratic trajectory.
The next two years are the critical window. The question is not whether institutions are under pressure — that is established — but whether they can maintain functional independence under sustained pressure. The specific tests:
The historical lag between institutional erosion and sovereign yield repricing is 3–12 years. Applied to the US timeline, this suggests a repricing window of 2028–2037. The specific mechanism: as democratic quality deteriorates, sovereign risk premia should increase, the yield curve should steepen on the long end, and the spread between US Treasuries, and other AAA sovereigns should widen.
The complication is that the US dollar's reserve currency status provides a "democratic discount" that no other country in the dataset enjoys. Treasuries are held not because investors trust American institutions but because there is no alternative asset of comparable depth and liquidity. This structural advantage could delay repricing by years or decades — or it could create a brittle equilibrium that snaps rather than bends when confidence finally shifts.
The US dollar's share of global foreign exchange reserves has declined from 72% in 2000 to approximately 58% in 2025. This decline predates the current democratic stress test and reflects structural diversification by central banks rather than a crisis of confidence. But the rate of decline matters. A gradual shift from 58% to 50% over a decade is a managed transition. A sharp drop below 50% in 2–3 years would signal that the institutional erosion has crossed a threshold where foreign central banks no longer view US governance as sufficiently stable for reserve asset status.
The BRICS+ payment system development, the expansion of bilateral currency swap agreements, and the increasing use of the Chinese yuan in commodity transactions are all incremental signals. None individually is decisive. Together, they describe an environment in which the exorbitant privilege that has underwritten American fiscal policy for 80 years is becoming, slowly but measurably, less exorbitant.
The single most important variable is the simplest: is the decline slowing down or speeding up? The PTI's 2-year window shows acceleration. The 10-year averages show a less alarming but still historically unusual velocity. The question is which timescale is more informative.
If the post-2024 acceleration was a step function — a one-time institutional shock that will be followed by sstabilisation at a lower level — then the 10-year averages are more informative, the event horizon has not been crossed, and the recovery probability is high. If the acceleration is the beginning of a positive-feedback loop — institutional weakening enabling further institutional weakening — then the 2-year window is the leading indicator, and the situation is considerably more serious.
The next 12 months of data will disambiguate between these two scenarios. A deceleration from −18.0/yr (2-year) towards the −3.0/yr range would suggest a step function. A sustained velocity above −10.0/yr would suggest a feedback loop. The stakes of this measurement are difficult to overstate.
The United States is exceptional in almost every relevant dimension. It is the world's oldest continuous democracy. It has the deepest institutional architecture. It is the wealthiest large country in history. Its military is institutionally apolitical. Its federal structure provides redundancy that no other backsliding country has possessed. Its civil society is dense, resourced, and combative.
And it is declining on every serious measure of democratic quality at a rate that has no precedent in the dataset of established democracies.
These two facts are not in tension. They are the tension. The question that the American exception poses to comparative politics is whether the structural advantages of a mature, wealthy democracy can provide a floor under democratic quality that the cross-national model does not capture — or whether those advantages merely raise the height from which the fall occurs.
The honest answer, as of February 2026, is that we do not know. The recalibration table shows that the conclusions depend critically on which index you trust, and that the credible range spans from "critical instability" to "stressed but intact." What the table also shows is that the direction of travel is not in dispute. Every index, every methodology, every measurement window agrees: the trajectory is down, and the velocity is historically unusual.
The American exception is not that it cannot happen here. The American exception is that it has never happened to a country with these structural advantages. Whether those advantages are sufficient is the question the next decade will answer.
This stress test was designed not to provide a single answer but to map the range of credible answers and identify the indicators that will disambiguate between them. The institutional tests of 2026–2028, the bond market repricing timeline, the dollar reserve trajectory, and above all the velocity of decline — these are the variables that will determine whether the American exception is a reassuring structural advantage or an untested hypothesis.
Watch them carefully.