Running the API

Democratic backsliding is not a switch, as if a country is either “a democracy” or “an autocracy,” and nothing meaningful happens in between. History tells a different story. What erodes first is not the form of government, but the tolerance for norm-breaking. Power tests boundaries long before it breaks constitutions.

That’s what the Autocracy Pressure Index (API) is meant to capture. Rather than going off feelings I've been running it for a few months now and refining it.

API doesn’t ask whether the United States is authoritarian. It asks how much structural pressure is being applied to democratic norms at a given moment—through institutions, information systems, security apparatuses, and incentives that reward coercive shortcuts. It’s a stress test, not a verdict.

When you back-cast this pressure biannually over the last forty years, the curve is revealing:

 Autocracy Pressure Index (API) — Biannual Estimates (1986–2026)

1986 2.2 Late–Cold War liberal consensus; strong institutions; low polarization

1988 2.3 Minor Iran-Contra aftershocks; institutions still resilient

1990 2.1 Post-Cold-War optimism; democratic confidence peak

1992 2.0 High trust, strong press norms

1994 2.4 Gingrich era begins norm hardening

1996 2.6 Media commercialization pressure begins

1998 2.9 Clinton impeachment → institutional weaponization

2000 3.3 Bush v. Gore → election legitimacy stress

2002 3.8 Post-9/11 security expansion (PATRIOT Act)

2004 4.1 Surveillance normalized; wartime executive latitude

2006 4.0 Iraq War fatigue; institutional pushback begins

2008 3.9 Financial crisis stress, but democratic norms intact

2010 4.2 Tea Party era; polarization accelerates

2012 4.4 Filibuster abuse, media fragmentation

2014 4.7 Disinformation ecosystems mature

2016 5.3 Election legitimacy + norm-breaking rhetoric

2018 5.6 Institutional stress tests; courts, press attacked

2020 6.1 Pandemic emergency powers + election denial

2021 6.5 Jan 6 aftermath; normalization of coercive rhetoric

2022 6.2 Partial institutional recovery; polarization locked

2023 6.4 Surveillance, executive immunity debates

2024 6.6 Election administration stress

2025 6.7 Knowledge warfare + institutional pressure

2026 7.0 Current run: pressure high [*see below appendix for details]

In the late 1980s and early 1990s, the U.S. sits around 2.0–2.3. Cold War consensus, high trust in institutions, strong press norms. The system is elastic; stress snaps back. Even Iran-Contra doesn’t meaningfully deform the material.

That changes in the late 1990s. By 1998–2000, the index rises to 2.9–3.3. Clinton’s impeachment and Bush v. Gore mark a shift: institutions are no longer just referees, but weapons. Democratic legitimacy absorbs its first visible microfractures.

After 9/11, the baseline never returns. From 2002–2006, API jumps to 3.8–4.1. This isn’t dictatorship—it’s security expansion. Surveillance and coercive capacity increase permanently. Security doesn’t equal autocracy, but it raises the floor on pressure.

The next bend comes after 2014. By 2016, the index hits 5.3, then 5.6 in 2018. This isn’t just polarization; it’s information collapse. Disinformation ecosystems mature. Knowledge itself becomes unstable. Every other pressure variable—courts, elections, civil society—gets amplified.

COVID and its aftermath push the system further. 2020–2021 reach 6.1–6.5: emergency powers, election denial, normalization of coercive rhetoric. Since then, the index stabilizes high—6.6 to 6.8 through 2026. Not collapse. Plastic deformation. Microfractures propagating.

This is why the question isn’t “Are we a dictatorship?” It’s whether we’re becoming comfortable living under sustained pressure.

The return of Donald Trump to the center of American politics isn’t the cause of this curve—it’s data within it. Attacks on courts, press, election administration, and executive accountability are not aberrations. They are stress tests applied to a system already under load.

Models don’t predict destiny. They reveal risk. And this one is flashing amber, not green.

[appendix, compares scores to Jan 4, 2026]

Component scores (0–10)
F — Press & speech
6.9 (↑ slight)
Sustained adversarial press environment
Security + foreign policy framing remains elevated
No new legal clampdown, but rhetoric + pressure persist
 +0.1 vs Jan 4

I — Institutional independence / rule of law
7.4 (↑)
Venezuela action still unresolved legally
Greenland tariff threats reinforce executive-unilateralism pattern
Congressional / judicial constraint signals remain weak or delayed
 +0.2 vs Jan 4

C — Civil society & assembly
6.3 (≈ flat)
CIVICUS “Obstructed” still the governing context
No new nationwide protest crackdown or expansion
Elevated, but not accelerating
 +0.2 vs Jan 4 (structural, not event-driven)

R — Repression / security apparatus
6.2 (↑)
Venezuela normalization effect persists
Greenland/NATO posture keeps coercive capacity salient
Still far from “police state,” but normalization continues
 +0.4 vs Jan 4 (largest proportional shift here)

M — Manipulation of elections / democratic façade
7.5 (≈ slight ↓ from peak)
No new election-administration shock
Still high due to legitimacy warfare + external conflict logic
Slight cooling from Jan 4 peak
 −0.1 vs Jan 4

S — Surveillance / digital control
6.2 (↑ slight)
Security rhetoric sustains tolerance
No statutory jump, but baseline remains elevated
 +0.1 vs Jan 4

E — Economic levers / capture
7.1 (↑)
Greenland tariff coercion now dominant driver
This is a clean API signal: economic tools used as political enforcement
This is the biggest change since your Jan 4 run
 +0.6 vs Jan 4

K — Knowledge / disinfo / propaganda
7.6 (↑)
Venezuela + Greenland narratives fully merged into partisan info warfare
Alliance fracture rhetoric intensifies narrative asymmetry
Structural amplifier still climbing
 +0.5 vs Jan 4 

API = F + I + C + R + M + S + E + K = 55.2/8 = 6.9