Polisci Regimes State Capacity Corruption Ir And Field Problems

---
name: polisci-regimes-state-capacity-corruption-ir-and-field-problems
description: "Use for comparative politics, international relations and the field's self-assessment: regimes and transitions including democratization and backsliding, state capacity, corruption and what anti-corruption interventions achieve, international relations theory, conflict, security and cooperation, the policy process, and the field's own problems with replication, measurement validity and contested data."
---

# Political Science and Governance: Regimes and Transitions, State Capacity, Corruption, IR Theory, Conflict and Cooperation, the Policy Process, and the Field's Own Problems

> **Part 5 of 6** of the *Political Science, Human Systems and Governance* reference (plugin `political-science-human-systems-governance`), covering §19–§25. Sibling skills: `polisci-what-the-field-establishes-power-the-state-and-theory` (§0–§4), `polisci-collective-action-commons-social-choice-and-voting` (§5–§8), `polisci-constitutional-design-electoral-systems-and-institutions` (§9–§13), `polisci-bureaucracy-parties-voter-behaviour-and-polarization` (§14–§18), `polisci-reference` (§26–§31). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.
>
> **Currency:** The theory is stable. Two areas are live. See §26 → `polisci-reference` for current democracy measurement, and the scholarly dispute about whether those measurements are reliable.

> **⚠️ A note on how this document handles contested questions.** ⚠️ **Political science
> contains robust empirical findings, contested empirical findings, and normative
> disputes that evidence cannot settle. This document tries to say which is which** (§1 → `polisci-what-the-field-establishes-power-the-state-and-theory`),
> **and where a question is genuinely contested it presents the strongest version of more
> than one position rather than picking.** ⚠️ **Where I report a finding, I attribute it;
> where scholars disagree, I say so.**
>
> **Complements a civil/industrial engineering reference (institutions as accumulated
> failure logs), a pedagogy reference (evidence quality in social science), and an
> economics-adjacent treatment of collective action.**
>
> **⚠️ GOTCHA** boxes mark the results that are counterintuitive or routinely misstated.
>
> **The three ideas that organize this document:**
> 1. **⚠️ COLLECTIVE ACTION PROBLEMS are the discipline's deepest structure** (§5–§6 → `polisci-collective-action-commons-social-choice-and-voting`).
>    **Individually rational behaviour producing collectively bad outcomes explains
>    government, taxation, regulation, unions, cartels and treaties — and Ostrom's work
>    showed the standard remedies are not the only ones.**
> 2. **⚠️ INSTITUTIONS SHAPE BEHAVIOUR, and the effects are often perverse** (§9–§15 → `polisci-constitutional-design-electoral-systems-and-institutions`, `polisci-bureaucracy-parties-voter-behaviour-and-polarization`).
>    **Electoral rules, veto points and federal structures produce predictable
>    consequences that designers frequently did not intend.**
> 3. **⚠️ MEASUREMENT IS THE FIELD'S BINDING CONSTRAINT** (§25, §26 → `polisci-reference`). **Most interesting
>    political concepts — democracy, power, legitimacy, polarization — are hard to measure,
>    and disputes that look substantive are often measurement disputes underneath.**

---

## §19. Regimes and Transitions

**⚠️ Regime typologies**: ⚠️ **liberal democracy, electoral democracy, COMPETITIVE
AUTHORITARIANISM (Levitsky & Way — real elections that are systematically unfair),
electoral authoritarianism, closed autocracy.** ⚠️ **The intermediate categories are where
most of the interesting cases and most of the measurement disputes live** (§26 → `polisci-reference`).
**⚠️ Democratization theories, competing rather than settled**: ⚠️ **modernization
(Lipset — wealth and democracy correlate; ⚠️ Przeworski's influential refinement is that
wealth may not CAUSE transitions but strongly predicts that democracies SURVIVE);
elite-pact transitions (O'Donnell & Schmitter); structural class accounts; and
international diffusion.**
**⚠️ Huntington's "waves"** framing remains a common organizing device.
**⚠️ BACKSLIDING has a characteristic modern form** that scholars broadly agree on
descriptively: ⚠️ **incremental and legal rather than a coup — executive aggrandizement,
capture of courts and electoral administration, pressure on media, and use of ordinary law
against opponents.** ⚠️ **Whether and how much this is happening in specific cases is
exactly what §26 → `polisci-reference` disputes.**
**⚠️ Authoritarian durability**: ⚠️ **the selectorate/winning coalition framework, and the
finding that party-based autocracies tend to outlast personalist ones.**

---

## §20. State Capacity

**⚠️ Arguably more explanatory than regime type for many outcomes.**
**⚠️ Dimensions**: ⚠️ **extractive (can it tax?), coercive, administrative, legal, and
infrastructural power (Mann — the ability to actually penetrate and implement across
territory, as distinct from despotic power over people).**
**⚠️ Measurement is genuinely difficult** — ⚠️ **tax/GDP ratios, census quality, birth
registration completeness and statistical capacity are all used as proxies, and each has
obvious problems.**
**⚠️ The institutions-and-development literature** — ⚠️ **Acemoglu, Johnson and Robinson on
inclusive versus extractive institutions and the colonial-origins instrument; North on
institutions as the rules of the game.** ⚠️ **Influential and heavily debated, with
critiques concerning the instrument's validity, measurement, and the direction of
causation.**
**⚠️ Weberian bureaucracy correlates with growth** (Evans & Rauch), ⚠️ **specifically
meritocratic recruitment and predictable careers.**
**⚠️ The resource curse** is reasonably supported: ⚠️ **states funded by resource rents
depend less on taxing citizens, which weakens the fiscal bargain that historically drove
accountability** (§3 → `polisci-what-the-field-establishes-power-the-state-and-theory`).

---

## §21. Corruption

**⚠️ Definitions vary** — ⚠️ **"abuse of public office for private gain" is standard and
contested, because it presupposes a public/private boundary that differs across societies.**
**⚠️ Types**: **grand vs petty; ⚠️ state capture (rewriting the rules) versus
administrative corruption (bending them); and ⚠️ "legal corruption" — lobbying, revolving
doors and campaign finance arrangements that are lawful and function similarly.**
**⚠️ Measurement caution**: ⚠️ **the best-known indices are PERCEPTION-based, which means
they can move with news coverage rather than behaviour and may reflect the reporter's
expectations.** **⚠️ Experience-based surveys and audit studies are methodologically
stronger but narrower in coverage.** ⚠️ **This is the same subjective-measure problem as
§26 → `polisci-reference`.**
**⚠️ What the evidence supports as anti-corruption measures**: ⚠️ **audits with publicized
results, e-government reducing discretionary contact, procurement transparency, and civil
service pay and meritocratic recruitment.** ⚠️ **Anti-corruption agencies have a mixed
record and can be captured or used selectively against opponents.**

---

# PART V — INTERNATIONAL

## §22. IR Theory

**⚠️ Presented as competing lenses, each capturing something real:**
```
⚠️ REALISM  anarchy (no world government) → self-help → security
   competition. ⚠️ Classical (human nature) vs NEOREALISM (structure,
   Waltz) vs offensive/defensive variants
   ⚠️ THE SECURITY DILEMMA is realism's best contribution: defensive
   measures are indistinguishable from offensive preparation, so
   both sides arm while both prefer peace — ⚠️ a collective action
   problem (§5) at the international level
⚠️ LIBERALISM  ⚠️ institutions, trade interdependence and domestic
   regime type change incentives. ⚠️ Institutions reduce
   transaction costs and provide information rather than
   overriding power
⚠️ CONSTRUCTIVISM  ⚠️ interests and identities are socially
   constructed, not given. "Anarchy is what states make of it"
   (Wendt). ⚠️ Explains normative change realism struggles with
MARXIST / world-systems · ⚠️ FEMINIST IR · ⚠️ POSTCOLONIAL IR
```
**⚠️ The DEMOCRATIC PEACE is among the more robust empirical regularities in IR**:
⚠️ **established democracies very rarely fight wars against each other.** ⚠️ **The
DISPUTE is over the mechanism — normative (shared norms of compromise) versus
institutional (accountability and audience costs) — and over definitional edge cases and
whether alliance structures during the Cold War confound it.**

---

## §23. Conflict, Security and Cooperation

**⚠️ Bargaining theory of war (Fearon)** is the standard framework: ⚠️ **because war is
costly, a negotiated settlement both sides prefer generally exists — so war requires an
explanation.** **⚠️ The three rationalist explanations are PRIVATE INFORMATION with
incentives to misrepresent, COMMITMENT PROBLEMS (neither side can credibly promise to
honour a deal, especially under shifting power), and ISSUE INDIVISIBILITY.**
**⚠️ Civil war** research: ⚠️ **the "greed versus grievance" debate has largely given way to
emphasis on state weakness, terrain, and the feasibility of insurgency — with the honest
caveat that these findings are sensitive to coding choices.**
**⚠️ Deterrence and nuclear strategy**: ⚠️ **credibility, second-strike capability, and the
stability-instability paradox — nuclear stability at the top may enable more conventional
conflict below it.**
**⚠️ Peacekeeping** — ⚠️ **the evidence is more favourable than public impressions suggest,
with several studies finding UN peacekeeping reduces the probability of conflict
recurrence, though selection effects are a genuine analytical difficulty.**

---

## §24. The Policy Process

```
⚠️ THE STAGES MODEL  agenda setting → formulation → adoption →
   implementation → evaluation. ⚠️ Useful teaching device;
   ⚠️ criticized as descriptively false — real policy is
   iterative, overlapping and messy
⚠️ MULTIPLE STREAMS (Kingdon)  ⚠️ problems, policies and politics
   flow separately; a POLICY WINDOW opens when they couple, and
   entrepreneurs exploit it. ⚠️ Explains why solutions often
   precede the problems they get attached to
⚠️ PUNCTUATED EQUILIBRIUM (Baumgartner & Jones)  ⚠️ long stability
   interrupted by rapid change when issue framing shifts
⚠️ ADVOCACY COALITION FRAMEWORK  coalitions bound by shared beliefs
⚠️ INCREMENTALISM (Lindblom)  ⚠️ "muddling through" as both a
   description and a defence — small steps are correctable
⚠️ IMPLEMENTATION  ⚠️ Pressman & Wildavsky's central finding:
   each link in a long chain of required agreements has a
   probability below 1, and the joint probability collapses
```
**⚠️ Evidence-based policy** — ⚠️ **and its honest limits: external validity (does a trial in
one context transfer?), value questions that evidence cannot settle, and the political
reality that evidence enters an arena rather than deciding it.**

---

## §25. ⚠️ The Field's Own Problems

> **⚠️ Stated plainly, because a reference that only reports findings without their
> reliability is misleading.**
```
⚠️ CAUSAL INFERENCE  ⚠️ no randomization at the country level (§1);
   the credibility revolution improved things via natural
   experiments, RDD and DiD, ⚠️ at the cost of studying what
   is identifiable rather than what matters most
⚠️ REPLICATION  ⚠️ political science has had its own replication
   and data-access reckoning, including high-profile fraud cases.
   ⚠️ Preregistration and results-blind review are now more common
⚠️ MEASUREMENT  ⚠️ the binding constraint (§26). Democracy, power,
   corruption, polarization, state capacity — all latent
   constructs measured by proxy or expert judgement
⚠️ SELECTION and SURVIVORSHIP in case selection
⚠️ WEIRD SAMPLES and a US-centric literature generalized abroad
⚠️ PREDICTION  ⚠️ the field's forecasting record on major events is
   poor. ⚠️ Tetlock's work found expert political judgement often
   barely beat simple baselines — while ALSO identifying that
   forecasting skill is real, trainable and unevenly distributed
⚠️ IDEOLOGICAL COMPOSITION  ⚠️ the professoriate is not
   representative of the public it studies, which is documented.
   ⚠️ Whether and how this biases findings is itself disputed,
   and it is a live question rather than a settled indictment
```