Structural mapping — external indices → pillars

Status: v0.2 (2026-04-28). First production mapping. Replaces the placeholder from iteration 3.

Principle

Each pillar (electoral, governance, judicial, media, civil, corruption) gets a 0–100 score as a weighted average of normalised values from selected external indicators. Indicators are chosen so that every pillar has ≥ 1 source, ideally ≥ 2 independent ones.

The mapping is explicit and fixed within a methodology version. Changes require a bump in methodology/CHANGELOG.md and recomputation of the historical series.

External indices used (Q2 2026)

Index	Year	CZ value	Original scale	Normalised 0–100	Source
V-Dem Liberal Democracy Index	2024	0.817	0–1	81.7	v-dem.net
EIU Democracy Index	2024	8.08	0–10	80.8	EIU 2024 PDF
Freedom House FitW	2025	95 (PR 37, CL 58)	0–100 (40+60)	95	freedomhouse.org/country/czechia/freedom-world/2025
RSF Press Freedom Index	2025	83.96 (rank 10)	0–100	83.96	rsf.org
TI Corruption Perceptions Index	2024	59 (rank 39)	0–100	59	transparency.org
WJP Rule of Law Index	2024	0.74 (rank 20/140)	0–1	74	worldjusticeproject.org

FH FitW per-category (2025)

Freedom House publishes 7 categories with their own scores. These are used in the mapping instead of the aggregate where it makes sense:

Category	Value	Normalised
A — Electoral Process	4 / 4	100
B — Political Pluralism & Participation	3.75 / 4	93.75
C — Functioning of Government	3.67 / 4	91.75
D — Freedom of Expression & Belief	4 / 4	100
E — Associational & Organizational Rights	4 / 4	100
F — Rule of Law	3.75 / 4	93.75
G — Personal Autonomy & Individual Rights	3.75 / 4	93.75

Mapping per pillar

Each pillar = unweighted mean of the chosen indices. Choice of indices per pillar:

electoral (weight 15 %)

Index	Value
V-Dem LDI	81.7
FH A — Electoral Process	100
FH B — Political Pluralism	93.75

Pillar score: (81.7 + 100 + 93.75) / 3 = 91.8

Rationale: V-Dem LDI is broader than just electoral, but it contains the EDI (Electoral Democracy Index) as a core component. FH A and B are direct measurements of elections and political competition.

governance (weight 20 %)

Index	Value
EIU Democracy Index	80.8
FH C — Functioning of Government	91.75

Pillar score: (80.8 + 91.75) / 2 = 86.3

Rationale: EIU has a separate "Functioning of government" category as part of its index, FH C measures the same thing directly. WJP "Constraints on Government Powers" would belong here, but we don’t have per-factor data; in iter 5 we use only these two.

judicial (weight 20 %)

Index	Value
WJP Rule of Law (overall)	74
FH F — Rule of Law	93.75

Pillar score: (74 + 93.75) / 2 = 83.9

Rationale: WJP overall is the leading international yardstick of judicial independence and law enforcement. FH F has a broader take on rule of law (covering police, protection from arrest). Without WJP per-factor data (Constraints on Government, Civil Justice, Criminal Justice), WJP overall is an acceptable proxy.

media (weight 15 %)

Index	Value
RSF Press Freedom Index	83.96
FH D — Freedom of Expression & Belief	100

Pillar score: (83.96 + 100) / 2 = 92.0

Rationale: RSF is the reference index for media freedom. FH D covers freedom of expression generally (not just media). The discrepancy between RSF (84) and FH D (100) is realistic — RSF picks up specific issues (SLAPPs, oligarchic concentration), FH D is more formal.

civil (weight 15 %)

Index	Value
FH E — Associational & Organizational Rights	100
FH G — Personal Autonomy & Individual Rights	93.75

Pillar score: (100 + 93.75) / 2 = 96.9

Rationale: civil liberties are best covered by FH subscores E (NGOs, unions, association) and G (personal autonomy, minority protection). The V-Dem civil liberties index would also fit, but we don’t have per-factor data.

corruption (weight 15 %)

Index	Value
TI Corruption Perceptions Index	59

Pillar score: 59.0

Rationale: TI CPI is the reference international yardstick for perceived corruption. WJP "Absence of Corruption" should be added, but we don’t have per-factor data; in the current iteration corruption is the only pillar with a single source. This is deliberate — TI CPI is the most respected and CZ-specific signal.

Consequence: the corruption pillar (59) is significantly lower than the others (84–97), reflecting a real discrepancy in Czechia — strong formal institutions but persistent perceived corruption. This does not need to be normalised, it is a faithful signal.

Weighted overall score

overall = electoral × 0.15 + governance × 0.20 + judicial × 0.20
        + media × 0.15 + civil × 0.15 + corruption × 0.15

       = 91.8 × 0.15 + 86.3 × 0.20 + 83.9 × 0.20
       + 92.0 × 0.15 + 96.9 × 0.15 + 59.0 × 0.15
       = 13.77 + 17.26 + 16.78 + 13.80 + 14.535 + 8.85
       = 85.0

Structural baseline 2026-Q2 overall: 85.0

For comparison: EIU 2024 places CZ at 80.8, V-Dem LDI at 81.7. Our aggregate of 85.0 is 3–4 points higher — reflecting the inclusion of FH (95) and weighting towards areas where CZ scores strongly (electoral, civil). The discrepancy is within the "normal variability between indices" range (~5–10 points), not a signal of bad mapping.

Update rules

1. Annual: when new reports come out (typically February–May), a new quarterly snapshot data/structural/{nextQ}.json is created with refreshed values. The mapping stays the same.
2. On a mapping change: bump version in methodology/CHANGELOG.md (v0.2 → v0.3), new quarterly snapshot, do not touch historical snapshots.
3. When per-factor data becomes available: if V-Dem or WJP publishes per-factor scores, the mapping can be refined — explicit changelog entry.

Open questions (sources for iter 6+)

1. WJP per-factor scores. WJP Rule of Law publishes 8 factors (Constraints on Govt, Absence of Corruption, Open Govt, Fundamental Rights, Order & Security, Regulatory Enforcement, Civil Justice, Criminal Justice). Per-factor data is in the interactive WJP profile or the Excel dataset — requires a one-off manual extraction, not a web scrape. When added:

   • governance adds WJP Constraints on Government + Regulatory Enforcement
   • judicial is refined with Civil Justice + Criminal Justice (drop the overall proxy)
   • corruption adds WJP Absence of Corruption (more than just TI CPI)
   • civil adds WJP Fundamental Rights + Order & Security

2. V-Dem per-component. V-Dem publishes dozens of components (v2x_polyarchy, v2x_libdem, v2x_partipdem, v2x_delibdem, v2x_egaldem, plus their sub-indices). For a future iteration, electoral should pull v2x_polyarchy, governance v2x_libdem (constraints), media v2x_freexp_altinf, etc.

3. Bertelsmann BTI mentioned in CLAUDE.md is not used in this iter — no freely available API/JSON, requires manual extraction. Iter 6+.

4. EC Rule of Law Report (CZ chapter) mentioned in CLAUDE.md is also not used — it contains qualitative assessment, not a numeric score. It could serve as a qualitative validator (e.g. if the EC report from year Y describes a specific problem in the judiciary, we expect a lower WJP score in Y+1).

5. Backtesting historical values. Iter 9+ should compute the baseline for 2018–2020 with the same methodology and compare with the EIU/V-Dem scores from that time.