Pre-registration
Maastricht convergence criteria 1992 imposed fiscal discipline that produced lower inflation and interest-rate convergence in pre-accession EU members, consistent with the Ordoliberal principle of rules-binding monetary constitutions.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
PRIMARY (dispositive): a difference-in-differences on the absolute CPI inflation gap to the Bundesbank anchor (DEU). Treated periphery = {ITA, ESP, PRT, GRC, IRL}; non-Eurozone OECD controls = {GBR, USA, JPN, CAN}. Pre-Maastricht window 1985-1991, post-Maastricht convergence window 1996-2002. For each country, gap_closure = |post_inflation - DEU_post| - |pre_inflation - DEU_pre|. PRIMARY = mean(treated gap_closure) - mean(control gap_closure). SUPPORTED if PRIMARY <= -3.0pp (treated converged on DEU at least 3pp more than controls). REFUTED if PRIMARY > -1.0pp (no relative convergence or wrong direction). Partial range: (-3.0, -1.0]pp. INFORMATIVE: per-country gap closures; absolute level of treated mean inflation in 1996-2002 (close-to-DEU implies absolute-not-just-relative convergence); government debt %GDP trajectories. METHOD_VALID: at least 4 of 5 treated and 3 of 4 control countries have WDI FP.CPI.TOTL.ZG observations in BOTH windows; if not, emit inconclusive.
formal test & threshold
test: maastricht_inflation_gap_did_treated_vs_controls threshold: PRIMARY: DiD <= -3.0pp for SUPPORTED. -3.0pp < DiD <= -1.0pp for partial. DiD > -1.0pp for refuted. METHOD_VALID gate: >= 4 treated, >= 3 control with both-window data.
Method
- Template
event_study- Clustering
country- Sample
- 13 countries · 1985 – 2002
- Evidence type
- associational
Two-window difference-in-differences on the absolute inflation gap to the Bundesbank anchor (DEU). Pre window 1985-1991 (pre-Maastricht); post window 1996-2002 (post-Maastricht convergence-test, ending at physical euro adoption). Per-country gap closure = |post_inflation - DEU_post| - |pre_inflation - DEU_pre|. Primary statistic = mean(treated_closure) - mean(control_closure). 10y sovereign yields are dropped from the dispositive test because vintage long-run yield series for ITA/ESP/PRT/GRC across 1985-1991 are not unified on disk; yields could be added to a v2 informative arm if eurostat:irt_lt_mcby and equivalents are pinned.
Data
| Variable | Source | Transform |
|---|---|---|
cpi_inflation outcome | world_bank_wdi:FP.CPI.TOTL.ZGtier 2 eurostat:prc_hicp_manrtier 1 ecb:ICPtier 1 | pct_change_yoy |
sovereign_yield_10y outcome | ecb:financial_markets_yields_10yrtier 1 fred:IRLTLT01tier 1 | level |
inflation_dispersion_eu outcome | eurostat:prc_hicp_manrtier 1 | cross_sectional_stddev |
maastricht_signature_event treatment | constructed:event date = 1992-02-07 (Maastricht Treaty signature); secondary = 1995-12 (Madrid summit naming euro)tier 5 | event_date |
pre_accession_eu_indicator treatment | constructed:binary = 1 for EU pre-accession members during 1992-2002tier 5 | binary |
government_debt_pct_gdp control | imf:GGXWDG_NGDPtier 2 eurostat:gov_10dd_edpt1tier 1 | level |
real_gdp_growth control | world_bank_wdi:NY.GDP.MKTP.KD.ZGtier 2 | pct_change_yoy |
pre_treatment_inflation control | world_bank_wdi:FP.CPI.TOTL.ZGtier 2 | lag_avg_1985_1991 |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Maastricht convergence: discipline effect on inflation
Verdict: SUPPORTED — Treated peripheral states (ITA/ESP/PRT/GRC/IRL) narrowed their inflation gap to the German anchor by 5.81pp from 1985-1991 to 1996-2002, vs only 1.32pp for non-Eurozone controls (GBR/USA/JPN/CAN). DiD = -4.49pp (<= -3.0pp dispositive threshold). Mean treated inflation fell from 9.47% to 3.20%.
Summary
- Pre-Maastricht window: 1985-1991.
- Post-Maastricht / convergence-test window: 1996-2002.
- Anchor (Bundesbank): DEU. Pre 1.85%, post 1.39%.
- Treated periphery (ITA/ESP/PRT/GRC/IRL): mean inflation 9.47% -> 3.20%, mean gap-closure to DEU -5.81pp.
- Non-Eurozone controls (GBR/USA/JPN/CAN): mean inflation 3.95% -> 1.53%, mean gap-closure to DEU -1.32pp.
- DiD (primary statistic): -4.49pp. Threshold for SUPPORTED: <= -3.0pp; partial range: (-3.0, -1.0]; refuted: > -1.0pp.
Per-country results
Treated periphery:
| Country | Pre infl. | Post infl. | |gap to DEU| pre -> post | Closure | |---|---|---|---|---| | ITA | 6.26% | 2.49% | +4.40pp -> +1.10pp | -3.30pp | | ESP | 6.73% | 2.82% | +4.88pp -> +1.43pp | -3.45pp | | PRT | 12.81% | 3.02% | +10.96pp -> +1.63pp | -9.33pp | | GRC | 17.97% | 4.47% | +16.12pp -> +3.08pp | -13.04pp | | IRL | 3.59% | 3.20% | +1.74pp -> +1.81pp | +0.07pp |
Controls:
| Country | Pre infl. | Post infl. | |gap to DEU| pre -> post | Closure | |---|---|---|---|---| | GBR | 5.58% | 1.84% | +3.73pp -> +0.45pp | -3.28pp | | USA | 3.95% | 2.40% | +2.09pp -> +1.01pp | -1.08pp | | JPN | 1.72% | -0.02% | +0.14pp -> +1.41pp | +1.27pp | | CAN | 4.56% | 1.92% | +2.71pp -> +0.53pp | -2.18pp |
Method
Difference-in-differences on the absolute inflation gap to the Bundesbank anchor (DEU). For each country we compute window-mean WDI FP.CPI.TOTL.ZG over the 1985-1991 (pre-Maastricht) and 1996-2002 (post-Maastricht convergence-test) periods, then the absolute gap to DEU's window-mean. Per-country gap closure = |post-gap| - |pre-gap| (negative = convergence). Primary statistic is the difference of group-mean closures (treated - control).
Yields convergence is informative-only: vintage 10y sovereign yields for ITA/ESP/PRT/GRC across 1985-1991 are not unified on disk; falsification is on inflation alone.
Data
- world_bank_wdi:FP.CPI.TOTL.ZG (CPI inflation, %YoY)
- imf:GGXWDG_NGDP (general government debt %GDP, informative)
Notes
Origin is auto-generated coverage-gap stub seeded from ordoliberal framing of Maastricht as rules-binding monetary constitution producing convergence. Human review required before promotion.